Guided Infrabony Review Willey Blackwell

Guided tissue regeneration for periodontal infra-bony defects (Review)
Needleman I, Worthington HV, Giedrys-Leeper E, Tucker R
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2008, Issue 4 http://www.thecochranelibrary.com
Guided tissue regeneration for periodontal infra-bony defects (Review) Copyright 2008 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
TABLE OF CONTENTS HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AUTHORS CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 1.1. Comparison 1 GTR versus control, Outcome 1 Attachment gain (change). . . . . . Analysis 1.2. Comparison 1 GTR versus control, Outcome 2 Sites gaining less than 2 mm attachment. Analysis 1.3. Comparison 1 GTR versus control, Outcome 3 Probing pocket depth (change). . . . Analysis 1.4. Comparison 1 GTR versus control, Outcome 4 Recession (change from baseline). . . . WHATS NEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 2 3 3 3 5 12 14 15 15 18 29 29 31 32 33 33 34 34 34 34 34
[Intervention Review]
Guided tissue regeneration for periodontal infra-bony defects

Ian Needleman1 , Helen V Worthington2 , Elaine Giedrys-Leeper1 , Richard Tucker1
1
Unit of Periodontology, Division of Restorative Dental Sciences, UCL Eastman Dental Institute, London, UK. 2 Cochrane Oral Health Group, MANDEC, School of Dentistry, The University of Manchester, Manchester, UK
Contact address: Ian Needleman, Unit of Periodontology, Division of Restorative Dental Sciences, UCL Eastman Dental Institute, University College London (UCL), University of London, 256 Grays Inn Road, London, WC1X 8LD, UK. i.needleman@eastman.ucl.ac.uk. Editorial group: Cochrane Oral Health Group. Publication status and date: Edited (no change to conclusions), published in Issue 4, 2008. Review content assessed as up-to-date: 12 January 2006. Citation: Needleman I, Worthington HV, Giedrys-Leeper E, Tucker R. Guided tissue regeneration for periodontal infra-bony defects. Cochrane Database of Systematic Reviews 2006, Issue 2. Art. No.: CD001724. DOI: 10.1002/14651858.CD001724.pub2. Copyright 2008 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
ABSTRACT Background Conventional treatment of destructive periodontal (gum) disease arrests the disease but does not usually regain the bone support or connective tissue lost in the disease process. Guided tissue regeneration (GTR) is a surgical procedure that specically aims to regenerate the periodontal tissues when the disease is advanced and could overcome some of the limitations of conventional therapy. Objectives To assess the efcacy of GTR in the treatment of periodontal infra-bony defects measured against conventional surgery (open ap debridement (OFD)) and factors affecting outcomes. Search strategy We conducted an electronic search of the Cochrane Oral Health Group Trials Register, MEDLINE and EMBASE up to April 2004. Handsearching included Journal of Periodontology, Journal of Clinical Periodontology, Journal of Periodontal Research and bibliographies of all relevant papers and review articles up to April 2004. In addition, we contacted experts/groups/companies involved in surgical research to nd other trials or unpublished material or to clarify ambiguous or missing data and posted requests for data on two periodontal electronic discussion groups. Selection criteria Randomised, controlled trials (RCTs) of at least 12 months duration comparing guided tissue regeneration (with or without graft materials) with open ap debridement for the treatment of periodontal infra-bony defects. Furcation involvements and studies specically treating aggressive periodontitis were excluded. Data collection and analysis Screening of possible studies and data extraction was conducted independently. The methodological quality of studies was assessed in duplicate using individual components and agreement determined by Kappa scores. Methodological quality was used in sensitivity analyses to test the robustness of the conclusions. The Cochrane Collaboration statistical guidelines were followed and the results expressed as mean differences (MD and 95% CI) for continuous outcomes and risk ratios (RR and 95% CI) for dichotomous outcomes calculated using random-effects models. Any heterogeneity was investigated. The primary outcome measure was change in clinical attachment.
Guided tissue regeneration for periodontal infra-bony defects (Review) Copyright 2008 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 1
Main results The search produced 626 titles, of these 596 were clearly not relevant to the review. The full text of 32 studies of possible relevance was obtained and 15 studies were excluded. Therefore 17 RCTs were included in this review, 16 studies testing GTR alone and two testing GTR + bone substitutes (one study had both test treatment arms). No tooth loss was reported in any study although these data are incomplete where patient follow up was not complete. For attachment level change, the mean difference between GTR and OFD was 1.22 mm (95% CI Random Effects: 0.80 to 1.64, Chi2 for heterogeneity 69.1 (df = 15), P < 0.001, I2 = 78%) and for GTR + bone substitutes was 1.25 mm (95% CI 0.89 to 1.61, Chi2 for heterogeneity 0.01 (df = 1), P = 0.91). GTR showed a signicant benet when comparing the numbers of sites failing to gain 2 mm attachment with risk ratio 0.54 (95% CI Random Effects: 0.31 to 0.96, Chi2 for heterogeneity 8.9 (df = 5), P = 0.11). The number needed to treat (NNT) for GTR to achieve one extra site gaining 2 mm or more attachment over open ap debridement was therefore 8 (95% CI 5 to 33), based on an incidence of 28% of sites in the control group failing to gain 2 mm or more of attachment. For baseline incidences in the range of the control groups of 3% and 55% the NNTs are 71 and 4. Probing depth reduction was greater for GTR than OFD: 1.21 mm (95% CI 0.53 to 1.88, Chi2 for heterogeneity 62.9 (df = 10), P < 0.001, I2 = 84%) or GTR + bone substitutes, weighted mean difference 1.24 mm (95% CI 0.89 to 1.59, Chi2 for heterogeneity 0.03 (df = 1), P = 0.85). For gingival recession, a statistically signicant difference between GTR and open ap debridement controls was evident (mean difference 0.26 mm (95% CI Random Effects: 0.08, 0.43, Chi2 for heterogeneity 2.7 (df = 8), P = 0.95), with a greater change in recession from baseline for the control group. Regarding hard tissue probing at surgical re-entry, a statistically signicant greater gain was found for GTR compared with open ap debridement. This amounted to a weighted mean difference of 1.39 mm (95% CI 1.08 to 1.71, Chi2 for heterogeneity 0.85 (df = 2), P = 0.65). For GTR + bone substitutes the difference was greater, with mean difference 3.37 mm (95% CI 3.14 to 3.61). Adverse effects were generally minor although with an increased treatment time for GTR. Exposure of the barrier membrane was frequently reported with a lack of evidence of an effect on healing. Authors conclusions GTR has a greater effect on probing measures of periodontal treatment than open ap debridement, including improved attachment gain, reduced pocket depth, less increase in gingival recession and more gain in hard tissue probing at re-entry surgery. However there is marked variability between studies and the clinical relevance of these changes is unknown. As a result, it is difcult to draw general conclusions about the clinical benet of GTR. Whilst there is evidence that GTR can demonstrate a signicant improvement over conventional open ap surgery, the factors affecting outcomes are unclear from the literature and these might include study conduct issues such as bias. Therefore, patients and health professionals need to consider the predictability of the technique compared with other methods of treatment before making nal decisions on use. Since trial reports were often incomplete, we recommend that future trials should follow the CONSORT statement both in their conduct and reporting. There is therefore little value in future research repeating simple, small efcacy studies. The priority should be to identify factors associated with improved outcomes as well as investigating outcomes relevant to patients. Types of research might include large observational studies to generate hypotheses for testing in clinical trials, qualitative studies on patient-centred outcomes and trials exploring innovative analytic methods such as multilevel modelling. Open ap surgery should remain the control comparison in these studies.
PLAIN LANGUAGE SUMMARY Guided tissue regeneration for periodontal infra-bony defects Current treatments for destructive periodontal (gum) disease are not able to restore damaged bone and connective tissue support for teeth. There are therefore limitations in treating patients with advanced disease. The surgical technique, guided tissue regeneration (GTR) may be able to achieve regeneration and therefore improve upon conventional surgical results. The results of this review have shown some advantage to using GTR in infra-bony defects but with wide variations in the benets achievable compared with conventional surgery. We were unable to identify conclusively factors responsible for this variability. Therefore, patients and health professionals need to consider the predictability of the technique compared with other methods of treatment before making nal
decisions on use. Adverse effects of treatment were generally minor and similar between groups although with an increased treatment time for GTR. We recommend further research to address the issue of variability and to identify which characteristics of the disease or the patient are more clearly associated with a benecial outcome.
BACKGROUND
Chronic periodontitis
Chronic periodontitis (CP) is a destructive gum condition, which is estimated to affect 10 to 30% of the world-wide population ( Baelum 1986; Le 1986; Oliver 1991). CP is caused by the bacteria within dental plaque, stimulating inammation within the periodontal tissues. In the susceptible individual, the result will be a breakdown of both the connective tissues which attach to the tooth and the supporting bone around the root. This usually results in the formation of a periodontal pocket around the root which acts as a reservoir for bacteria. The morbidity of this condition can be underestimated and may include an uncomfortable loosening of the teeth (which may impair eating), cosmetic problems (as teeth drift or gums recede), a tendency to abscess formation within the pocket and eventual tooth loss.
Treatment of periodontitis
The objectives for treating periodontitis are mainly concerned with stabilising or arresting the condition and the crucial role of the patients home care plaque control is well recognised (Lindhe 1975). The debridement of bacterial deposits coating the surface of the root, deep within the periodontal pocket is also essential and is achieved in the rst instance, by scaling techniques. In addition, periodontal surgery is used where the depth of the deposits within the pocket, prevents adequate access for debridement.
treatment of deep defects and are a risk factor for further deterioration (Claffey 1990). A particular concern for many patients is that conventional surgery tends to increase gum recession, which can cause cosmetic problems. In an attempt to overcome some of these limitations, surgical techniques have been developed to regenerate the tissues lost in the disease process. The most documented of these is guided tissue regeneration (GTR) and case reports have shown healing with the formation of new attachment (Garrett 1996). In this procedure, a biocompatible barrier membrane (either resorbable or non-resorbable) is surgically implanted to cover and protect the bone defect. If non-resorbable, the barrier is surgically removed 4 to 6 weeks after implantation. Connective tissue and bone regeneration may then occur within the bone defect protected by the barrier. The biological explanation for this procedure is the prevention of migration of the epithelial periodontal tissues into the crater, allowing time for bone and other attachment tissues to heal. During normal healing, it appears that the epithelial tissues migrate rapidly into the wound, preventing regeneration (Karring 1993). However understanding of this process is incomplete. Although GTR is accepted into clinical practice, a systematic review of its effectiveness is important since the procedure is technically demanding and nancially costly, in view of the expense of the barrier membranes (approximately US$100 to $300).
OBJECTIVES
The objective of this review is to assess the effect of GTR in the treatment of periodontal infra-bony defects measured against conventional access ap surgery (open ap debridement), with respect to clinical, radiographic and patient-centred outcomes.
The infra-bony defect and its treatment

Often, the bony destruction in periodontitis forms a crater-like defect around the root and this is called an infra-bony defect when the base of the periodontal pocket projects into and is surrounded by the jaw bone. Infra-bony defects present major challenges to periodontal treatment. Following treatment, healing occurs by repair but without the formation of new supporting tissue (Caton 1976). In addition, residual periodontal pockets may remain after
METHODS
Criteria for considering studies for this review

3
Types of studies Randomised, controlled trials (RCTs) of at least 12 months duration. Quality assessment issues are addressed in the methods of review.
Search methods for identication of studies

A search of the Cochrane Oral Health Group Trials Register, EMBASE and MEDLINE was conducted up to and including April 2004. We used the strategy: guided tissue regeneration OR guided-tissue-regeneration OR GTR OR periodontal regeneration OR periodontal-regeneration OR intra bony defect* OR intrabony defect* OR intra-bony defect* OR infra bony defect* OR infrabony defect* OR infra-bony defect* OR intra osseous OR intraosseous OR intra-osseous. For MEDLINE this also included the rst two phases of the sensitive search strategy for randomised controlled trials of the Cochrane Oral Health group and was adapted for other databases. Handsearching included a complete search of Journal of Periodontology, Journal of Clinical Periodontology and Journal of Periodontal Research up to April 2004 and bibliographies of all relevant papers and review articles. In addition, for the original review we contacted experts/groups/companies involved in surgical research to nd other trials or unpublished material or to clarify ambiguous or missing data and posted requests for data on two periodontal electronic discussion groups: International Association for Dental Research - Dental Faculty, Periodontal Research Group website (www.iadr-dentalfaculty.org/sigs/index.html) and the Periodont electronic newsletter (periodont@spallek.com) up to September 2000.
Types of participants Patients with a clinical diagnosis of chronic periodontitis (CP) or periodontitis in subjects aged 21 years or older. Studies that address GTR exclusively around furcation involved teeth were excluded. The lower age limit was selected in order to be as inclusive of studies as possible but studies specically treating aggressive periodontitis were excluded as this is currently classied as a different form of periodontitis (much less common and of more rapid progression) and might respond differently to treatment (Tonetti 1999).
Types of interventions (1) Guided tissue regeneration (GTR) versus open ap debridement. (2) GTR and bone substitutes combined versus open ap debridement.
Types of outcome measures
Data collection and analysis

Titles and abstracts of the search results, were screened by two independent review authors (Richard Tucker (RT), Ian Needleman (IN)). The full text of all studies of possible relevance were obtained for independent assessment by three review authors (RT, IN and Elaine Giedrys-Leeper ( EGL)) against the stated inclusion criteria. Any disagreement was resolved by discussion amongst the review authors and authors of the trials were contacted to provide missing data where possible. Data entry on to a computer was performed by two review authors (RT and Helen Worthington (HW)).
Primary outcomes
Tooth loss Change in attachment levels Patient well-being or quality of life.
Secondary outcomes
Methodological quality (protection from bias) Change in probing depths Change in gingival recession Changes in bone/hard tissue (a) Radiographic, (b) Hard tissue probing at surgical re-entry Disease recurrence (% sites with 2 mm loss of probing attachment measured from 12 months after treatment) Percentage of sites with 4 mm probing depth at completion of study Aesthetics (change: better or worse in patients opinion) Post-operative complications (including pain, infection) Economic outcomes. The methodological quality of included studies was assessed using components shown to affect study outcomes (typically by exaggeration of treatment effect) including, allocation concealment (concealment of the randomisation code from those recruiting patients to avoid selection bias) and blinding of examiners and therapists (Juni 2001; Moher 1998; Schulz 1995). Methodological quality was used in sensitivity analyses to test the robustness of the conclusions but was not used to exclude studies qualifying for the review on the basis of their inclusion criteria. Agreement between investigators in the review for these components was checked using Kappa statistics.
4
Concealment of allocation criteria Adequate allocation concealment would include either central randomisation or sequentially numbered sealed opaque envelopes. Inadequate allocation concealment if there is an open allocation sequence and the participants and trialists can foresee the upcoming assignment. Unclear allocation concealment where the method cannot be determined. Blinding of examiner criteria This assessed whether persons measuring the outcome of care were aware of which treatment the participant received, and was graded as yes, no or unclear. Blinding of surgeon criteria This could only be achieved if defect preparation was complete prior to informing the operator of treatment allocation. For splitmouth studies, this would demand preparation of both test and control sites simultaneously and was graded as yes, no or unclear. Data synthesis A weighted treatment effect was calculated and the results were expressed as mean differences (MD) and 95% condence interval (CI) for continuous outcomes and risk ratio (RR) and 95% CI for dichotomous outcomes. The analysis for the continuous outcome variables was conducted using the generic inverse variance statistical method where the mean differences and standard errors are entered for all studies so that the parallel group studies and intraindividual (split-mouth) studies could be combined (see Results). Variance imputation methods were used to estimate appropriate variance estimates in some split-mouth studies, where the appropriate standard deviation of the differences was not included in study reports (Follmann 1992). The primary outcome measure was gain in attachment. Random-effects models were used for analyses and the signicance of discrepancies in the estimates of the treatment effects from the different trials were assessed by means of Cochrans test for heterogeneity. If any signicant heterogeneity (P < 0.1) was detected it was planned to investigate this. Publication bias was examined using both the Begg and Mazumdar rank correlation test and the Egger regression asymmetry test (Begg 1994; Egger 1997). A subgroup analysis was conducted for the eight parallel group and eight split-mouth studies for attachment change to see whether it is appropriate to combine these study designs in the meta-analysis. Only one of the ve split-mouth studies (Cortellini 1998) stated the standard deviation of the mean difference for two of the outcomes (change in attachment and change in probing depth). It was possible to calculate this for these two outcomes from the raw data
in another split-mouth study (Chung 1990) and to estimate this from P-values in a further study (Mora 1996). The intra-patient correlations from these studies ranged from 0.14 to 0.43. An intra-patient correlation of 0.25 was used throughout this review to estimate the standard errors for the other ve split-mouth studies (Blumenthal 1990; Loos 2002; Pontoriero 1999; Pritlove-Carson 1995; Ratka-Kruger 2000). Sensitivity analyses were conducted imputing standard errors for the intra-patient correlation of zero, which would provide a very conservative estimate of the standard error. Subgroup analyses were planned to investigate the effects of factors thought to be most inuential on periodontal outcomes including smoking status, barrier membrane type (resorbable versus non-resorbable) and surgical technique (conventional ap versus papilla preservation ap). A further sensitivity analysis examined periodontal outcomes with exclusion of studies providing more frequent post-operative maintenance than is practicable in clinical practice (dened as maintenance visits more frequent than every 3 months, commencing 3 months post-surgery).
RESULTS
Description of studies
See: Characteristics of included studies; Characteristics of excluded studies. See Characteristics of included studies table. Of the 17 included trials (Table 1), three were multicentre studies with the following numbers of individuals completing each study; 136 patients in a parallel group design (Tonetti 1998), 23 patients in an intra-individual design (Cortellini 1998) and 109 patients employing a parallel group design (Cortellini 2001). Of the single-centre trials, there were ve intra-individual (splitmouth) trials ranging from 10 to 40 participants (Blumenthal 1990; Chung 1990; Mora 1996; Pontoriero 1999; Pritlove-Carson 1995), seven parallel group trials ranging from 36 to 90 participants (Cortellini 1995; Cortellini 1996; Kim 1998; Mayeld 1998; Sculean 2001; Silvestri 2000; Zucchelli 2002) and two studies containing both split-mouth and parallel group elements (Loos 2002; Ratka-Kruger 2000) contributing 16 and 25 individuals respectively. Six trials were based in private practice (Cortellini 1995; Cortellini 1996; Cortellini 1998; Cortellini 2001; Sculean 2001; Tonetti 1998), nine trials were Universitybased (Blumenthal 1990; Chung 1990; Kim 1998; Loos 2002; Mayeld 1998; Mora 1996; Pontoriero 1999; Pritlove-Carson 1995; Ratka-Kruger 2000) and for two trials, the setting was not clear (Silvestri 2000; Zucchelli 2002). Three studies (Chung 1990; Mayeld 1998; Tonetti 1998) were supported, in part, by companies whose products were being used as interventions in the trials.
5
Table 1. Summary of meta-analyses for different outcomes and sensitivity analyses Outcome Study type N of studies 16 Het. Chi2 69.1 34.4 Het. P-value <0.001 <0.001 Mean difference 1.22 1.71 95% CI (RE) 0.80, 1.64 1.02, 2.40
Attachment gain GTR only
Attachment gain GTR parallel 8 group studies Attachment gain GTR split- 8 mouth studies Attachment gain GTR + bone 2 substitutes Attachment gain Sensitivity analy- 16 sis: GTR only assuming intra-patient correlation =0 Attachment gain Sensitivity analy- 7 sis: GTR only adequate allocation concealment Attachment gain Sensitivity anal- 8 ysis: GTR only, examiner blind Attachment gain Sensitivity anal- 8 ysis: GTR only, therapist blind Attachment gain Sensitivity anal- 3 ysis: GTR only, examiner and therapist blind Probing depth GTR only reduction 11
18.5
0.01
0.79
0.37, 1.21
0.01
0.91
1.25
0.89, 1.61
63.3
<0.001
1.24
0.82, 1.66
30.4
<0.001
1.63
0.85, 2.42
44.8
<0.001
1.27
0.50, 2.04
32.4
<0.001
1.19
0.57, 1.81
4.4
0.11
0.41
-0.33, 1.08
62.9
<0.001
1.21
0.53, 1.88
Probing depth GTR parallel 5 reduction group studies Probing depth GTR split- 6 reduction mouth studies Probing depth GTR + bone 2 reduction substitutes
43.6
<0.001
1.59
0.21, 2.97
9.1
0.11
0.87
0.38, 1.36
0.03
0.85
1.24
0.89, 1.59
Table 1. Summary of meta-analyses for different outcomes and sensitivity analyses
(Continued)
Gingival recession Gingival recession Gingival recession Gingival recession
GTR only
2.6
0.92
0.26
0.08, 0.44
GTR parallel 4 group studies GTR split- 4 mouth studies GTR + bone 1 substitutes 3
0.97
0.81
0.15
-0.12, 0.42
0.41
0.94
0.35
0.11, 0.60
N/A
N/A
-0.33
-0.43, -0.23
Bone gain. Sur- GTR only gical re-entry
0.85
0.65
1.39
1.08, 1.71
Bone gain. Sur- GTR + bone 1 gical re-entry substitutes
N/A
N/A
3.37
3.14, 3.61
The age range of all the included studies was from 21 to 81. However, two studies presented mean ages of 48.7 to 50.8 years ( Mayeld 1998) and 36 to 39 years (Loos 2002) . All studies had both male and female participants, but with differing proportions. One study had more participants who were smokers than nonsmokers, with 21 out of 38 smokers (Mayeld 1998). See Characteristics of excluded studies table. There were 15 excluded studies seven of which reported 6 months data only (Chaves 1996; Eickholz 1996; Kilic 1997; Kim 1996; Kwan 1998; Nygaard-Ostby 1996; Quteish 1992). Of the other eight studies, six were written in English, one in German and the other in Japanese. The latter two were translated revealing that one (Eger 1998) did not have access ap control and the other ( Yoshinari 1996) was not an RCT. Of the remaining studies, three were either not fully randomised controlled trials (Becker 1996; Eickholz 1998; Zybutz 2000) one was quasi-randomised (Bratthall 1998) and a further study did not employ GTR (Shamiri 1992). One used a unique radiographic assessment for which the outcomes could not be compared with other radiographic techniques (Iversen 1996).
Risk of bias in included studies

Randomisation was reported in all studies included in this review. An adequate method of randomisation, was present in the following studies (Blumenthal 1990; Cortellini 1995; Cortellini 1996; Cortellini 1998; Cortellini 2001; Loos 2002; Mora 1996; Pritlove-
Carson 1995; Sculean 2001; Tonetti 1998; Zucchelli 2002). It was unclear as to how the following studies randomised their treatment groups (Chung 1990; Kim 1998; Mayeld 1998; Pontoriero 1999; Ratka-Kruger 2000; Silvestri 2000) Concealment of the randomisation code during patient selection, was adequate in nine studies (Cortellini 1995; Cortellini 1996; Cortellini 1998; Cortellini 2001; Mora 1996; PritloveCarson 1995; Sculean 2001;Tonetti 1998; Zucchelli 2002) and this was not clear in seven studies (Blumenthal 1990; Chung 1990; Kim 1998; Loos 2002; Mayeld 1998; Pontoriero 1999; Silvestri 2000). Examiner blinding was described in eight of the studies ( Cortellini 1995; Cortellini 1996; Kim 1998; Loos 2002; Mayeld 1998; Pontoriero 1999; Sculean 2001; Zucchelli 2002). Operator blinding (such as by revealing the treatment code only after defect preparation was complete) was present in the following eight studies (Cortellini 1995; Cortellini 1996; Cortellini 1998; Loos 2002; Pontoriero 1999; Pritlove-Carson 1995; Sculean 2001; Tonetti 1998). In all studies, participants entering a study could be accounted for at study completion. However, a high number of withdrawals or drop outs were reported for two studies; 33% drop outs (Chung 1990) 35% losses (Kim 1998) and much smaller numbers for the multicentre studies (Cortellini 1998; Cortellini 2001; Tonetti 1998).
Effects of interventions
Inter-investigator agreement The Kappa scores (standard error (SE)) for key elements of interinvestigator agreement were the following: allocation concealment 0.64 (0.26), therapist blinding 0.64 (0.26) and examiner blinding 0.88 (0.24). These scores indicate a good level of agreement. Tooth loss No teeth were reported lost in either experimental group in any study. However, in studies where patient follow up was incomplete, it was unclear whether tooth loss might have contributed to this decit. Comparisons
Change in attachment level (Comparison 1, Outcomes 1.1, 1.2)
A subgroup analysis was conducted for the eight parallel group and eight split-mouth studies for attachment change to see whether it is appropriate to combine these study designs in the metaanalysis (Subgroups 1.1.1 and 1.1.2). Only one of the ve splitmouth studies (Cortellini 1998) stated the standard deviation of the mean difference for two of the outcomes (change in attachment and change in probing depth). It was possible to calculate this for these two outcomes from the raw data in another splitmouth study (Chung 1990) and to estimate this from P-values in a further study (Mora 1996). The intra-patient correlations from these studies ranged from 0.14 to 0.43. An intra-patient correlation of 0.25 was used throughout this review to estimate the standard errors for the other ve split-mouth studies (Blumenthal 1990; Loos 2002; Pontoriero 1999; Pritlove-Carson 1995; RatkaKruger 2000). Sensitivity analyses were conducted imputing standard errors for the intra-patient correlation of zero, which would provide a very conservative estimate of the standard error. Sixteen studies presented attachment level data for GTR alone, eight parallel group trials (Cortellini 1995; Cortellini 1996; Cortellini 2001; Mayeld 1998; Sculean 2001; Silvestri 2000; Tonetti 1998; Zucchelli 2002) and eight split-mouth studies ( Blumenthal 1990; Chung 1990; Cortellini 1998; Loos 2002; Mora 1996; Pontoriero 1999; Pritlove-Carson 1995; RatkaKruger 2000) The results for this analysis show a statistically signicantly greater attachment gain for test groups compared with open ap debridement. For GTR the weighted mean difference between test and control was 1.22 mm (95% CI Random Effects
0.80 to 1.64, Chi2 for heterogeneity 69.1 (df = 15), P < 0.001, I2 = 78%) There appeared to be a difference between the two subgroups based on the study design with the subgroup of splitmouth studies having a lower treatment effect 0.79 mm (95% CI Random Effects 0.37 to 1.21, compared to that for the parallel group studies 1.71 mm (95% CI Random Effects 1.02 to 2.40). This difference was statistically signicant (metaregression slope coefcient -0.91, 95% CI -1.71 to -0.11, P = 0.026). The sensitivity analysis imputing an intra-patient correlation of zero only made a small difference to the estimate for the split-mouth studies, with weighted mean difference 0.82 mm (95% CI Random Effects 0.40 to 1.24). There were two studies for GTR + bone substitutes (Blumenthal 1990; Kim 1998) and the weighted mean difference for these two studies was 1.25 mm (95% CI 0.89 to 1.61, Chi2 for heterogeneity 0.01 (df = 1), P = 0.91), similar to the overall result for GTR alone. From an additional (post-hoc) analysis for GTR, in six parallel group studies (Cortellini 1995; Cortellini 1996; Cortellini 2001; Mayeld 1998; Tonetti 1998; Zucchelli 2002) it was possible to extract data for number of sites gaining less than 2 mm of attachment. This again showed a signicant benet for GTR, risk ratio 0.54 (95% CI Random Effects 0.31 to 0.96, Chi2 for heterogeneity 8.91 (df = 5), P = 0.11) (Comparison 1, Outcome 1.2). The number needed to treat (NNT) for GTR to achieve one extra site gaining 2 mm or more attachment over open ap debridement was therefore 8 (95% CI 5 to 33), based on an incidence of 28% of sites in the control group failing to gain 2 mm or more of attachment. For baseline incidences in the range of the control groups of 3% and 55% the NNTs are 71 and 4. It should be noted that the calculation of NNT values based on arbitrary cut-thresholds derived from continuous data may not be appropriate. We are currently examining this phenomenon.
Sensitivity analysis In order to investigate the robustness of the results with respect to very frequent maintenance that may be impractical to provide in many clinics (< 3 monthly after the rst 3 months postsurgery), we conducted a sensitivity analysis excluding seven studies with very frequent follow up (included studies: Blumenthal 1990; Cortellini 2001, Chung 1990; Loos 2002, Mora 1996; Pritlove-Carson 1995, Silvestri 2000, Ratka-Kruger 2000; Tonetti 1998. This showed a small reduction of the weighted mean difference to 0.95 mm (95% CI Random Effects 0.4960 to 1.40) which was not signicant (metaregression slope coefcient -0.62 (se = 0.43), P = 0.15) (Table 2).
Table 2. Random-effects metaregression analysis of attachment change
Characteristic
Number of studies
Slope estimate -0.62
95% CI -1.46 to 0.22
Slope interpretation
P-value
Frequency of main- 16 tenance
Higher attach- 0.15 ment change for visits < 3 months Higher attach- 0.11 ment change for non-absorbable membrane Higher attachment 0.09 change for papilla preservation
Type of GTR barrier 18 comparisons
-0.70
-1.55 to 0.15
Surgical technique
16
0.75
-0.10 to 1.59
A subgroup analysis was also conducted comparing the six studies with papilla preservation with the 10 studies where the papilla was not fully preserved. Although a higher attachment change was found for papilla preservation this subgroup analysis was not statistically signicant (P = 0.09, Table 2). Comparison of studies using an absorbable versus non-absorbable membrane type also gave rise to a non-signicant difference between the subgroups (P = 0.11; Table 2). Sensitivity analyses for attachment level change which excluded poorer quality studies resulted in reduced numbers of incorporated studies although heterogeneity remained statistically signicant ( Table 1). When studies without surgeon blinding were excluded, the summary estimate was similar and heterogeneity was still signicant. When studies without both surgeon blinding and examiner blinding were excluded, the difference between GTR and OFD became non-statistically signicant and with no signicant heterogeneity (three studies: mean difference 0.41mm 95% CI 0.33 to 1.08, heterogeneity P = 0.11). Caution should be exercised here as the number of studies in these comparisons was also much lower which in itself could reduce heterogeneity. Subgroup analysis of parallel group and split-mouth studies reduced the heterogeneity, although there was still a great deal of unexplained heterogeneity. However as all the ndings for each outcomes were in the same direction we felt it was appropriate to undertake the meta-analyses as shown.
Change in probing depth (Comparison 1, Outcome 1.3)
2002; Mora 1996; Pontoriero 1999; Ratka-Kruger 2000). The standard errors for one split-mouth study was given in the report (Cortellini 1998) and could be estimated for a further study ( Mora 1996). The intra-patient correlations were 0.11 and 0.1224 and so a value of 0.1 was used to calculate the standard errors for the other split-mouth studies. A sensitivity analysis was conducted imputing standard errors for the intra-patient correlation of zero, which would provide a very conservative estimate of the standard error. The results demonstrated a signicantly greater probing depth reduction for GTR, mean difference of 1.21 mm (95% CI 0.53 to 1.88, Chi2 for heterogeneity 62.9 (df = 10), P < 0.001, I2 = 84%). Although the treatment effect for split mouth studies was lower, this was not statistically signicant (metaregression slope coefcient -0.72, 95% CI -2.21 to 0.78, P = 0.35). There were also two studies for GTR + bone substitutes ( Blumenthal 1990; Kim 1998) with weighted mean difference 1.24 mm (95% CI 0.89 to 1.59, Chi2 for heterogeneity 0.03 (df = 1), P = 0.85) similar to that for the GTR alone.
Gingival recession (Comparison 1, Outcome 1.4)
There were 11 studies for GTR alone including change in probing depth as an outcome, ve parallel group studies (Cortellini 2001; Mayeld 1998; Sculean 2001; Silvestri 2000; Zucchelli 2002) and six split-mouth studies (Blumenthal 1990; Cortellini 1998; Loos
Nine studies for GTR had gingival recession as an outcome, four with a parallel group design (Cortellini 2001; Mayeld 1998; Sculean 2001; Zucchelli 2002) and ve with a split-mouth design (Blumenthal 1990; Loos 2002; Mora 1996; Pontoriero 1999; Ratka-Kruger 2000) and for GTR + bone substitutes one study (Blumenthal 1990). We were unable to estimate any of the intra-patient correlations and we decided to use a value of 0.25 for the split-mouth studies for this outcome. A statistically signicant difference between GTR and open ap debridement controls was
evident (weighted mean difference 0.26 mm (95% CI Random Effects 0.08 to 0.43, Chi2 for heterogeneity 2.7 (df = 8), P = 0.95), with a greater change in recession from baseline for the control group. The single study of GTR + bone substitutes showed slightly greater recession for test than controls, with mean difference -0.33 mm (95% CI -0.43 to -0.23). Bone change Radiographic data was contained in two studies. One used intra-oral radiographs and customised stents to aid reproducibility of radiographic geometry (Mayeld 1998). This showed 0.6 mm gain in bone from the base of the defect in both test and control groups. The other study reported using standardised radiographs (Ratka-Kruger 2000). Descriptive data only were reported indicating only minute changes in bone structure were noticeable. Regarding surgical re-entry, three studies (Blumenthal 1990; Chung 1990; Mora 1996) reported data for GTR alone and one study for the combination of GTR + bone substitute (Blumenthal 1990). For GTR, a statistically signicant greater gain in hard tissue probTable 3. Adverse events and patient reported outcomes Study Blumenthal 1990 Chung 1990 Cortellini 1995 Outcome
ing was found for GTR compared with open ap debridement. This amounted to a weighted mean difference of 1.39 mm (95% CI 1.08 to 1.71, Chi2 for heterogeneity 0.85 (df = 2), P = 0.65). For GTR + bone substitutes the difference was greater, with a mean difference 3.37 mm (95% CI 3.14 to 3.61).
Adverse effects and patient-reported outcomes There were only limited data on these outcomes (Table 3). Healing was generally reported to be uneventful for both test and control groups. Exposure of the GTR barrier membrane was widely reported and ranged from 20% of sites treated with the non-resorbable titanium-reinforced material (Cortellini 1995) to 68% for non-resorbable ePTFE barriers (Mayeld 1998). There seemed little difference on membrane exposure between the different materials, particularly comparing resorbable versus non-resorbable. The effect of membrane exposure was generally reported to be modest although such an event could result in the need for more rigorous maintenance (and therefore appointments) or the use of systemic antibiotics.
Throughout the study, No infection, untoward reactions, infection or delayed healing occured. During the post-surgical healing phase of this study, no adverse tissue reactions were noted in any test sites. Healing in all cases was uneventful. Membrane exposure occurred in 20% of cases in the test group [GTR titanium] and in 60% of cases in the GTR control group [GTR ePTFE]; in all these instances the extent of exposure was minimal and always limited to a small portion of the interdental tissue. All sites healed uneventfully. No clinically detectable or subjectively reported side effects were reported in any treated patient. No reported Pain: No signicant difference for postoperative pain between groups (VAS scale). Morbidity: 36% test patients reported that the procedure interfered with daily activity for an average of 2.7 days, 32% controls reported that procedure interfered with daily activity for an average of 2.4 days. Post-operative oedema was most frequent complication: GTR 62% patients, acces ap 40% (estimated from gure 3) statistically greater for GTR P = 0.01. GTR barrier membrane exposure was greatest at 3 weeks after surgery at 54% of membranes. Clinical healing was uneventful in both groups. Limited signs of inammation, swelling, or redness were observed.
Cortellini 1996
Cortellini 1998 Cortellini 2001
Kim 1998
10
Table 3. Adverse events and patient reported outcomes
(Continued)
Mayeld 1998
All sites healed uneventfully, and no clinical signs of infection or ap instability were observed. There were no side effects reported in any treated patient. Membrane exposure: by 3 weeks after surgery 15/22 (68%) had shown some exposure. Membrane exposure: 40% by 3 weeks after surgery. For these sites, maintenance was intensied (mouthwash and topical antiseptic). No suppuration observed. Not reported. All surgical sites healed with minimum patient discomfort and no signs of infection, with the exception of one case where a purulent discharge was observed on the outer aspect of one membrane around a lower molar at the fourth week. Antibiotics were prescribed and the membrane removed at 6 weeks. No serious wound healing problems or inammatory reactions were seen at the sites that had received membranes. Membrane exposure: 49% of sites by week 4. The postoperative healing was generally uneventful and consisted mainly of postoperative swelling and/or diarrhoea due to the antibiotic regimen. Neither allergic reactions nor suppuration were observed after any of the treatments. Membrane exposure: 6/14 (43%) sites No reported. Not reported. All sites healed uneventfully. Membrane exposure: 10/33 sites (33%). each outcomes were going in the same direction we felt it was appropriate to undertake the meta-analyses as shown.
Mora 1996
Pontoriero 1999 Pritlove-Carson 1995
Ratka-Kruger 2000
Sculean 2001
Silvestri 2000 Tonetti 1998 Zucchelli 2002
Time for procedure One study (Cortellini 2001) measured time taken for the procedures. The average surgical time for GTR procedures was 98.8 minutes (SD 45.7) and for access ap surgery 74.9 minutes (SD 33.6) and this was highly statistically signicant (P = 0.001).
Prognostic factors
Impact of study quality
Heterogeneity
Sensitivity analyses for attachment level change which excluded poorer quality studies resulted in reduced numbers of incorporated studies although heterogeneity remained statistically signicant (see Table 1). Although subgroup analysis of parallel group and split-mouth studies reduced the heterogeneity there was still a great deal of unexplained heterogeneity, however as all the ndings for
The variability in reporting data on prognostic factors such as initial defect depth, plaque levels and smoking prevented a meaningful comparison. One study (Mayeld 1998) presented a subgroup analysis comparing clinical changes in smokers and non-smokers. This showed reduced benets in smokers for attachment gain (GTR group: non-smokers 1.9 mm SD 1.5, smokers 0.8 mm SD 0.8) although with little effect on probing depth change. However, this result should be viewed with some caution as the groups were not intentionally balanced with respect to disease levels.
Publication bias
11
We tested publication bias using the Begg and Mazumdar rank correlation test, which found no evidence of a correlation between the effect estimates and their variances, indicating no evidence of publication bias (P = 0.53). The results of the Egger regression asymmetry test also suggested no evidence of publication bias (P = 0.43). There is no strong evidence of publication bias and 16 studies were included in this analysis.
protection from bias as well as factors affecting clinical heterogeneity.
Protection from bias

Since study quality has been shown to have a direct impact on the size of the effect of treatment we explored this effect with sensitivity analyses including allocation concealment, examiner and therapist blinding. The results of these analyses were not consistent. For attachment gain, excluding studies without adequate concealment of the randomisation code or examiner blinding did not substantially affect the estimate or reduce heterogeneity. However, when studies without both surgeon blinding and examiner blinding were excluded, the difference between GTR and OFD became non-statistically signicant and with no signicant heterogeneity (three studies: mean difference 0.41 mm 95% CI -0.33 to 1.08, heterogeneity P = 0.41). Caution should be exercised here as the number of studies in these comparisons was also much lower which in itself could reduce heterogeneity. However, the trend to a reduction of the magnitude of effect with greater protection from bias is in line with previous studies of the impact of bias (Juni 2001; Moher 1998; Schulz 1995). If further studies are to be conducted on GTR, it is critical that they employ greater methodological rigour and in particular in their protection against these biases. Publication bias was investigated and found not to be statistically signicant. Whilst these tests are conservative in their ability to demonstrate such bias, the number of studies included (16) should be adequate to identify publication bias if it was present. Therefore, we can conclude that there is no evidence of an effect of publication bias on the summary values.
Other outcomes
No data were found for the following outcomes: Disease recurrence (% sites with 2 mm loss of probing attachment measured from 12 months after treatment) Ease of maintenance (% of sites with < 4 mm probing depth) Aesthetics (change: better or worse in patients opinion) Cost/benet (treatment time plus estimated material costs).
DISCUSSION Guided tissue regeneration alone

This systematic review has shown an overall mean increase in attachment gain for GTR over open ap debridement (mean difference 1.22 mm, 95% CI 0.8 to 1.64). However, this value may not be a valid estimate of effect since the difference between study outcomes (heterogeneity) is substantial and statistically signicant. Therefore, the value should not be quoted to demonstrate the magnitude of difference between the two therapies. It should be noted that 11/16 trials produced a statistically signicant greater gain in attachment with GTR than OFD and with no statistically signicant publication bias (see below). Therefore, we suggest that the data indicate that GTR can produce a statistically signicant greater gain in attachment however, the magnitude of this superiority is not clear. Tooth loss values did not give evidence of a difference between studies. Since the studies were of no more than 12 months duration, such a nding is not surprising and may not be helpful in distinguishing between treatments. Progression of attachment loss may proceed at rates from 0.1 to 3mm per year (Cobb 1996; Lindhe 1989). Therefore, multi-year studies are likely to be needed to evaluate tooth loss adequately. The results of the 16 randomised controlled trials included in this review show a substantial variation in their results. The mean additional attachment gain from GTR over that achieved by open ap debridement surgery ranged between studies from 0.02 to 3.60 mm. This range is large and the differences are difcult to reconcile. We have attempted to explore some of the possible causes of this heterogeneity and these analyses should be viewed as exploratory observations. These investigations included analyses of
Clinical heterogeneity
Available data allowed the investigation of several clinical aspects that we hypothesised, a priori, could affect heterogeneity. These included, frequency of supportive maintenance care (more than every 3 months versus 3-monthly after the rst 3 months of healing), type of GTR barrier membrane (resorbable versus non-resorbable) and surgical technique (papilla preservation technique versus conventional approach). Random-effects metaregression analyses were conducted to compare each of these subgroups. The hypothesis for excluding studies with frequent maintenance was based on the nding that frequency of maintenance care can affect periodontal surgical results (Westfelt 1983). However, the analysis for this review showed neither a statistically signicant effect on the summary estimate nor an elimination of heterogeneity (P = 0.15). Membrane type might be important. The difference between GTR and OFD for attachment gain was greater for nonabsorbable barriers than absorbable barriers. However this difference was not statistically signicant by metaregression (P = 0.11).
12
Surgical techniques aiming to retain the interdental soft tissue have been proposed with the potential to achieve and maintain primary closure during wound healing. It has been suggested that such methods could therefore produce greater clinical improvements ( Cortellini 1999). The meta-analysis for the six studies with papilla preservation did not show a statistically signicant difference compared with the overall estimate despite the apparent greater attachment gain (0.75 mm, 95% CI -0.10 to 1.59, P = 0.09) and therefore the heterogeneity remained highly statistically signicant. However, the 95% condence interval only just includes a value of no difference suggesting that papilla preservation might be important and should be examined in future studies. Another explanation for the heterogeneity might be variability between studies in prognostic factors that have been documented to affect the outcome of regenerative surgery. These include; plaque levels, cigarette smoking and defect severity as expressed by baseline probing depth, attachment level and bone defect depth at baseline. Regarding plaque and smoking, it is apparent that differences in the way that both factors are reported between studies prevent sensible comparison. For instance, some studies present full mouth plaque scores, others measure plaque at the experimental sites only and other studies present plaque index values or no plaque data. The effect of smoking on reducing the gain in attachment following surgery is reported in only one subgroup analysis of an RCT (Mayeld 1998). This result highlights the need for more research into prognostic factors to help explain heterogeneity. Therefore, the extent to which we have been successful in explaining the troubling extent of the heterogeneity has been limited. It should be noted however, that heterogeneity existed not only between studies but also within them. In two large multicentre trials (Cortellini 2001; Tonetti 1998) the results between centres within each study showed a substantial variability (from best to worst) in attachment gain of between 1.73 mm and 2.1 mm respectively. Therefore, although efcacy of GTR has been demonstrated in some studies, the effectiveness and generalisability of such a technique has not been demonstrated. Variability of results is clearly an important issue when considering the relevance of a treatment to clinical practice. Although we have explored some of the possible causes of heterogeneity in this systematic review, we have been unable to determine denitively what factors account for this. The number and characteristics of studies currently available is insufcient to answer this clinically relevant problem. In terms of clinical signicance, mean difference values are difcult to interpret. However, the risk ratio for sites gaining less than 2 mm attachment demonstrated that sites treated with GTR were 38% less likely to fail to attain a gain of 2 mm of attachment or more than those treated by open ap debridement (risk ratio 0.54 (95% CI 0.31 to 0.96). There was no signicant heterogeneity between the studies. The number needed to treat (NNT) for GTR to achieve one extra site gaining 2 mm or more attachment over
open ap debridement was 8 (95% CI 5 to 33) i.e. 8 patients need to be treated for one to achieve this benet over OFD. This is unchanged from the original review despite the inclusion of data from two further studies (Cortellini 2001; Zucchelli 2002). Other clinical outcomes indicate statistically greater improvements with GTR compared with OFD. As with attachment level gain, heterogeneity with probing depth hampers a conclusion of the size of this improvement. The analysis of recession indicated more recession following the use of OFD compared with GTR and with no heterogeneity (0.26 mm, 95% CI 0.08 to 0.43, heterogeneity P = 0.96, n = 9). The result for gingival recession is interesting since although 7 of the 9 studies produced more recession with OFD, this difference was statistically signicant in only one study ( Pontoriero 1999). However, the greater precision that is achievable when multiple studies are combined with meta-analysis means that overall, statistically signicantly more recession can be shown to occur following OFD.
Adverse effects and patient-reported outcomes

Adverse effects were generally minor and similar between GTR and access ap surgery. Membrane exposure was a common nding and represents the main difference in postoperative healing between groups. Whilst membrane exposure has been associated with poorer clinical outcomes (Selvig 1992) one study in this series of trials did not nd a negative effect on clinical attachment level (Cortellini 2001). A further possible implication with membrane exposure is an increase in need for professional plaque removal visits or the provision of systemic antibiotics as reported in these studies, either for prevention or treatment of infection around the membrane. However, there is no clear evidence of how such a situation is best managed.
Effect of study design

This analysis has demonstrated a statistically signicant difference between parallel group and split-mouth studies with respect to attachment level change. Split-mouth studies produced a more conservative estimate of attachment level gain. Whilst there remained statistically signicant heterogeneity in both subgroups, the difference between split-mouth and parallel group was statistically signicant by metaregression. To our knowledge, this is the rst time that such a difference has been demonstrated and underlines the importance of analysing by study design. The reasons leading to smaller differences between two interventions can only be speculated upon. This might be a chance nding as a result of producing two subgroups of studies. Possibly, protection from bias could be more straightforward in split-mouth studies. For instance, selection bias might be at less risk as the patient provides both experimental groups. Furthermore, split-mouth studies might facilitate
13
the maintenance of masking for patient, examiner and therapist if the patient provides both groups. Alternatively it may be possible that there is some cross-over benet, either local or systemic as has been previously suggested, and this effect could reduce the difference in outcome between interventions (Hujoel 1992).
GTR plus bone grafts

Only two studies could be located examining the combination of GTR and bone grafts versus open ap debridement in a randomised, 12 month design (Blumenthal 1990; Kim 1998). The effects of the combination treatment were similar to GTR alone for attachment gain, but with slightly more probing depth reduction (1.2 mm), greater gain in hard tissue probing at re-entry surgery (3.4 mm) in one study and a gain in gingival recession in a further study.
Laurell 1998) and unclear selection criteria for randomised controlled trials in the second, including the inclusion of studies of shorter duration (16). A recent systematic review (Murphy 2003) on GTR produced broadly similar ndings to our present review. In terms of gain in clinical attachment, GTR produced 0.81 mm greater gain than OFD (P < 0.001) (no condence interval presented for the difference). The result contained highly statistically signicant heterogeneity. This group did not nd a statistically signicant difference between different types of barrier material. Differences in search strategy, inclusion of both randomised and non-randomised studies and of studies of shorter duration of follow up, may have accounted for these differences. The results of this update are similar to those of the original review, even though six more trials have been included. The major difference was found for gingival recession which is now signicant. This shows a greater increase in gingival recession from baseline with the control group of access ap. In addition, further exploration of heterogeneity has now been possible.
Reporting quality
We found that many study reports were incomplete in their presentation of methods or results. We contacted 19 authors to clarify missing or ambiguous data. However, we would recommend that authors of RCTs follow the CONSORT statement (www.consortstatement.org) which provide clear guidance on presentation of trial reports and would help systematic reviewing of the literature.
AUTHORS CONCLUSIONS Implications for practice

Eleven out of 16 studies showed greater attachment gain for guided tissue regeneration (GTR) over open ap debridement. However, this systematic review has shown that the outcomes following GTR are highly variable, both between and within studies. Therefore, patients and health professionals need to consider the predictability of the technique compared with other methods of treatment before making nal decisions on use. A meta-analysis comparing GTR with open ap debridement indicates greater clinical attachment gain of 1.22 mm (95% CI 0.80 to 1.64) and probing depth reduction of 1.21 mm (95% CI 0.53 to 1.88) for GTR over open ap debridement. However, the highly statistically signicant heterogeneity between studies indicates that these summary values may be not reliable summary values for the magnitude of probing changes. Statistically signicantly greater gingival recession occurs following access ap surgery than following the use of guided tissue regeneration (mean difference 0.26 mm 95% CI 0.08 to 0.43) and with no statistically signicant heterogeneity. There are few data from controlled studies of 12 months duration on the combination of bone substitutes with guided tissue regeneration but the results suggest little added benet beyond a gain in hard tissue probing at surgical re-entry. Few data exist to answer important questions such as patient evaluation of outcomes. The use of papilla preservation aps versus conventional ap designs should be further evaluated in randomised controlled trials.
14
Statistical methods
This review considered parallel group and split-mouth studies. In theory the combining of the treatment effects from these studies should be straightforward, however, due to the poor reporting of the data from the split-mouth studies, the standard deviation of the difference had to be estimated. This was achieved by using the results of the split-mouth studies that presented the necessary data to calculate the intra-patient correlation for the other split-mouth studies. Sensitivity analyses were carried out imputing different values for the intra-patient correlation and the results of these were very similar to the results presented in this review.
Comparison with previous thorough reviews and meta-analyses

Two recent meta-analyses have reported greater benets to GTR than found in the present systematic review. One review indicated a difference in attachment gain between GTR and open ap debridement of 2.7 mm (Laurell 1998) and a second review reported 1.6 mm difference (Cortellini 2000). Differences in the methods of these reviews that may help to explain the results include the incorporation of uncontrolled and unblinded studies in one (
Implications for research

There is little value in future research repeating, small efcacy studies. Studies of sufcient size to examine the effect of possible prognostic or technical factors on outcomes are needed. These might include, type of infra-bony defect, type of barrier membrane and type of surgical ap design. Types of research might include large observational studies to generate hypotheses for testing in clinical trials, qualitative studies on patient-centred outcomes and trials exploring innovative analytic methods such as multilevel modelling. Open ap surgery should remain the control comparison in these studies. Control of bias (especially randomisation, concealment of allocation and blinding of examiner and operator where possible) needs to be more rigorously employed in study designs and this might be an important factor in explaining heterogeneity. Greater attention in designing and reporting studies should be given to study quality issues as set out in the CONSORT statement. This will help to facilitate the evaluation of these studies. Greater consideration in study design should be given to outcomes that capture the boarder patient experience including patient centred and economic outcomes and evaluation of adverse effects. Two treatments that produce similar clinical changes may be quite different from the patient or economic perspective.
ACKNOWLEDGEMENTS
We would like to thank Sylvia Bickley at the Cochrane Oral Health Group (OHG) in Manchester UK, for her tremendous help in searching the literature and Emma Tavender, also at the Cochrane OHG for her administrative support. We are deeply grateful to the many researchers and clinicians who have provided clarication of data and comments on this review.
REFERENCES
References to studies included in this review

Blumenthal 1990 {published data only} Blumenthal N, Steinberg J. The use of collagen membrane barriers in conjunction with combined demineralized bone-collagen gel implants in human infrabony defects. Journal of Periodontology 1990;61(6):31927. Chung 1990 {published data only} Chung KM, Salkin LM, Stein MD, Freedman AL. Clinical evaluation of a biodegradable collagen membrane in guided tissue regeneration. Journal of Periodontology 1990;61(12):7326. Cortellini 1995 {published and unpublished data} Cortellini P, Pini Prato G, Tonetti M. Periodontal regeneration of human intrabony defects with titanium reinforced membranes. A
controlled clinical trial. Journal of Periodontology 1995;66(9): 797803. Cortellini 1996 {published and unpublished data} Cortellini P, Pini Prato G, Tonetti M. Periodontal regeneration of human intrabony defects with bioresorbable membranes. A controlled clinical trial. Journal of Periodontology 1996;67(3): 21723. Cortellini 1998 {published and unpublished data} Cortellini P, Carnevale G, Sanz M, Tonetti M. Treatment of deep and shallow intrabony defects. A multicenter randomized controlled clinical trial. Journal of Clinical Periodontology 1998;25 (12):9817.
15
Cortellini 2001 {published data only} Cortellini P, Tonetti MS, Lang NP, Suvan JE, Zuchelli G, Vangsted T, et al.The simplied papilla preservation ap in the regenerative treatment of deep intrabony defects: clinical outcomes and postoperative morbidity. Journal of Periodontology 2001;72(12): 170212. Kim 1998 {published data only} Kim CK, Chai JK, Cho KS, Moon IS, Choi SH, Sottosanti JS, et al.Periodontal repair in intrabony defects treated with a calcium sulfate implant and calcium sufate barrier. Journal of Periodontology 1998;69(12):131724. Loos 2002 {published data only} Loos BG, Louwerse PH, van Winkelhoff AJ, Burger W, Gilijamse M, Hart AA, et al.Use of barrier membranes and systemic antibiotics in the treatment of intraosseous defects. Journal of Clinical Periodontology 2002;29(10):91021. Mayeld 1998 {published data only} Mayeld L, Soderholm G, Hallstrom H, Kullendorff B, Edwardsson S, Bratthal G, et al.Guided tissue regeneration for the treatment of intraosseous defects using a bioabsorbable membrane. A controlled clinical study. Journal of Clinical Periodontology 1998; 25(7):58595. Mora 1996 {published and unpublished data} Mora F, Etienne D, Ouhayoun J. Treatment of interproximal angular defects by GTR: 1 year follow-up. Journal of Oral Rehabilitation 1996;23(9):599606. Pontoriero 1999 {published data only} Pontoriero R, Wennstrom J, Lindhe J. The use of barrier membranes and enamel matrix proteins in the treatment of angular bone defects. Journal of Clinical Periodontology 1999;26(12): 83340. Pritlove-Carson 1995 {published and unpublished data} Pritlove-Carson S, Palmer R, Floyd P. Evaluation of GTR in the treatment of paired periodontal defects. British Dental Journal 1995;179(10):38894. Ratka-Kruger 2000 {published data only} Ratka-Kruger P, Neukranz E, Raetzke P. Guided tissue regeneration procedure with bioresorbable membranes versus conventional ap surgery in the treatment of infrabony periodontal defects. Journal of Clinical Periodontology 2000;27(2):1207. Sculean 2001 {published data only} Sculean A, Windisch P, Chiantella GC, Donos N, Brecx M, Reich E. Treatment of intrabony defects with enamel matrix proteins and guided tissue regeneration. A prospective controlled clinical study. Journal of Clinical Periodontology 2001;28(5):397403. Silvestri 2000 {published data only} Silvestri M, Ricci G, Rasperini G, Sartori S, Cattaneo V. Comparison of treatment of intrabony defects with enamel matrix derivative, guided tisue regeneration with a nonresorbable membrane and modied Widman ap. A pilot study. Journal of Clinical Periodontology 2000;27(8):60310. Tonetti 1998 {published and unpublished data} Tonetti MS, Cortellini P, Suvan JE, Adriaens P, Baldi C, Dubravec D, et al.Generalizability of the added benets of guided tissue regeneration in the treatment of deep intrabony defects. Evaluation
in a multi-center randomized controlled clinical trial. Journal of Periodontology 1998;69(11):118392. Zucchelli 2002 {published data only} Zucchelli G, Bernardi F, De Motebugnoli L. Enamel matrix proteins and guided tissue regeneration with titanium-reinforced expanded polytetrauoroethylene membranes in the treatment of infrabony defects: a comparative controlled clinical trial. Journal of Periodontology 2002;73(1):312.
References to studies excluded from this review

Becker 1996 {published data only} Becker W, Becker BE, Mellonig J, Caffesse RG, Warrer K, Caton JG, et al.A prospective multi-center study evaluating periodontal regeneration for Class II furcation invasions and intrabony defects after treatment with a bioresorbable membrane: 1-year results. Journal of Periodontology 1996;67(7):6419. Bratthall 1998 {published data only} Bratthall G, Soderholm G, Neiderud AM, Kullendorff B, Edwardsson S, Attstrom R. Guided tissue regeneration in the treatment of human infrabony defects. Clinical, radiographical and microbiological results: a pilot study. Journal of Clinical Periodontology 1998;25(11):90814. Chaves 1996 {published data only} Chaves ES, Geurs NC, Reddy MS, Jeffcoat MK. Clinical and radiographic digital imaging evaluation of a bioresorbable membrane in the treatment of periodontal bone defects. International Journal of Periodontics and Restorative Dentistry 1996; 16(5):44353. Eger 1998 {published data only} Eger T, Muller H-P. [Parodontale Regeneration in vertikalen Konochendeften mit resorbierbaren Membranen und SchumelMatrix-Protein]. Deutsche Zahnarztliche Zeitschrift 1998;53:5904. Eickholz 1996 {published data only} Eickholz P, Benn DK, Staehle HJ. Radiographic evaluation of bone regeneration following periodontal surgery with or without expanded polytetrauoroethylene barriers. Journal of Periodontology 1996;67(4):37985. Eickholz 1998 {published data only} Eickholz P, Lenhard M, Benn DK, Staehle HJ. Periodontal surgery of vertical bony defects with or without synthetic bioabsorbable barriers. 12-month results. Journal of Periodontology 1998;69(11): 121017. Iversen 1996 {published data only} Iversen B, Albandar J, Oydna J, Gjermo P. Bone density after 1 year in periodontal lesions treated surgically with or without ePTFE membrane placement. Journal of Clinical Periodontology 1996;23 (6):5126. Kilic 1997 {published data only} Kilic AR, Efeoglu E, Yilmaz S. Guided tissue regeneration in conjunction with hydroxyapatite-collagen grafts for intrabony defects. A clinical and radiological evaluation. Journal of Clinical Periodontology 1997;24(6):37283. Kim 1996 {published data only} Kim CK, Choi EJ, Cho KS, Chai JK, Wikesjo UM. Periodontal repair in intrabony defects treated with a calcium carbonate implant
16
and guided tissue regeneration. Journal of Periodontology 1996;67 (12):13016. Kwan 1998 {published data only} Kwan SK, Lekovic V, Camargo PM, Klokkevold PR, Kenney EB, Nedic M, et al.The use of autogenous periosteal grafts as barriers for the treatment of intrabony defects in humans. Journal of Periodontology 1998;69(11):12039. Nygaard-Ostby 1996 {published data only} Nygaard-Ostby P, Tellefsen G, Sigurdsson TJ, Zimmerman GJ, Wikesjo UM. Periodontal healing following reconstructive surgery: effect of guided tissue regeneration. Journal of Clinical Periodontology 1996;23(12):10739. Quteish 1992 {published data only} Quteish D, Dolby AE. The use of irradiated-crosslinked human collagen membrane in guided tissue regeneration. Journal of Clinical Periodontology 1992;19(7):47684. Shamiri 1992 {published data only} Shahmiri S, Singh I, Stahl S. Clinical response to the use of the HTR Polymer Implant in human intrabony lesions. International Journal of Periodontics and Restorative Dentistry 1992;12(4):2949. Yoshinari 1996 {published data only} Yoshinari N, Tohya T, Inagaki K, Mori A, Nishiyama S, Koide M, et al.5 years of clinical evaluation of nonresorbable membranes in the treatment of intrabony defects following guided tissue regeneration. Nagoya, Japan: Department of Periodontology, School of Dentistry, Aichigakuin University 1996; Vol. 494:2119. Zybutz 2000 {published data only} Zybutz MD, Laurell L, Rapoport DA, Persson GR. Treatment of intrabony defects with resorbable materials, non-resorbable materials and ap debridement. Journal of Clinical Periodontology 2000;27(3):16978.
Cortellini 1999 Cortellini P, Prato GP, Tonetti MS. The simplied papilla preservation ap. A novel surgical approach for the management of soft tissues in regenerative procedures. International Journal of Periodontics and Restorative Dentistry 1999;19(6):58999. Cortellini 2000 Cortellini P, Tonetti M. Focus on intrabony defects: guided tissue regeneration. Periodontology 2000 2000;22:10432. Egger 1997 Egger M, Davey Smith G, Schneider M, Minder C. Bias in metaanalysis detected by a simple, graphical test. BMJ 1997;315(7109): 62934. Follmann 1992 Follmann D, Elliott P, Suh I, Cutler J. Variance imputation for overviews of clinical trials with continuous response. Journal of Clinical Epidemiology 1992;45(7):76973. Garrett 1996 Garrett S. Periodontal regeneration around natural teeth. Annals of Periodontology 1996;1(1):62166. Hujoel 1992 Hujoel PP, DeRouen TA. Validity issues in split-mouth trials. Journal of Clinical Periodontology 1992;19(9):6257. Juni 2001 Juni P, Altman DG, Egger M. Systematic reviews in health care: Assessing the quality of controlled clinical trials. BMJ 2001;323 (7303):426. Karring 1993 Karring S, Nyman T, Gottlow J, Laurell L. Development of the biological concept of guided tissue regeneration - animal and human studies. Periodontology 2000 1993;1:2635. Laurell 1998 Laurell L, Gottlow J, Zybutz M, Persson R. Treatment of intrabony defects by different surgical procedures. A literature review. Journal of Periodontology 1998;69(3):30313. Lindhe 1975 Lindhe J, Nyman S. The effect of plaque control and surgical pocket elimination on the establishment and maintenance of periodontal health. A longitudinal study of periodontal therapy in cases of advanced disease. Journal of Clinical Periodontology 1975;2 (2):6779. Lindhe 1989 Lindhe J, Okamoto H, Yoneyama T, Haffajee A, Socransky SS. Periodontal loser sites in untreated adult subjects. Journal of Clinical Periodontology 1989;16(10):6718. Le 1986 Le H, Anerud A, Boysen H, Morrison E. The natural history of periodontal disease in man. Rapid, moderate and no loss of attachment in Sri Lankan labourers 14 to 46 years of age. Journal of Clinical Periodontology 1986;13(5):43145. Moher 1998 Moher D, Pham B, Jones A, Cook DJ, Jadad AR, Moher M, et al.Does quality of reports of randomised trials affect estimates of intervention efcacy reported in meta-analyses?. Lancet 1998;352 (9128):60913.
17
Additional references
Baelum 1986 Baelum V, Fejerskov O, Karring T. Oral hygiene, gingivitis and periodontal breakdown in adult Tanzanians. Journal of Periodontal Research 1986;21(3):22132. Begg 1994 Begg CB, Mazumdar M. Operating charateristics of a rank correlation test for publication bias. Biometrics 1994;50(4): 1088101. Caton 1976 Caton J, Zander HA. Osseous repair of an infrabony pocket without new attachment of connective tissue. Journal of Clinical Periodontology 1976;3(1):548. Claffey 1990 Claffey N, Nylund K, Kinger T, Garret S, Egelberg J. Diagnostic predictability of scores of plaque, bleeding, suppuration and probing depth for probing attachment loss. Journal of Clinical Periodontology 1990;17(2):10814. Cobb 1996 Cobb CM. Non-surgical pocket therapy: mechanical. Annals of Periodontology 1996;1(1):44390.
Murphy 2003 Murphy KG, Gunsolley JC. Guided tissue regeneration for the treatment of periodontal intrabony and furcation defects. A systematic review. Annals of Periodontology 2003;8(1):266302. Oliver 1991 Oliver R, Brown L, Loe H. Variations in the prevalence and extent of periodontitis. Journal of the American Dental Association 1991; 122(6):438. Schulz 1995 Schulz K, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995;273 (5):40812. Selvig 1992 Selvig K, Kersten B, Chamberlain A, Wikesjo U, Nilveus R. Regenerative surgery of intrabony periodontal defects using ePTFE barrier membranes. Scanning electron microscopic evaluation of retrieved membranes vs. clinical healing. Journal of Periodontology 1992;63(12):9748. Tonetti 1998 Tonetti M, Cortellini P, Suvan JE, Adriaens P, Baldi C, Dubravec
D, et al.Generalizability of the added benets of guided tissue regeneration in the treatment of deep intrabony defects. Evaluation in a multi-center randomized controlled clinical trial. Journal of Periodontology 1998;69(11):118392. Tonetti 1999 Tonetti M, Mombelli A. Early-onset periodontitis. Annals of Periodontology 1999;4(1):3953. Westfelt 1983 Westfelt E, Nyman S, Socransky SS, Lindhe J. Signicance of frequency of professional tooth cleaning for healing following periodontal surgery. Journal of Clinical Periodontology 1983;10(2): 14856.
References to other published versions of this review

Needleman 2001 Needleman IG, Giedrys-Leeper E, Tucker RJ, Worthington HV. Guided tissue regeneration for periodontal infra-bony defects. Cochrane Database of Systematic Reviews 2001, Issue 2. [Art. No.: CD001724. DOI: 10.1002/14651858.CD001724] Indicates the major publication for the study
18
CHARACTERISTICS OF STUDIES
Characteristics of included studies [ordered by study ID]

Blumenthal 1990 Methods RCT, split mouth, 5 treatment groups, 12 months duration. 10 individuals, 4 female, aged 34-57, with moderate to advanced periodontal disease. Control: OFD Test 1: Resorbable collagen membrane. Test 2: Collagen membrane over combined AAA bone-collagen implant. Antigen-extracted allogenic bone implant (not included in this analysis). Combined AAA bone-collagen implant (not included in analysis). Maintenance intervals: 3 monthly after rst 2 months post-surgery. Hard tissue measurements at re-entry surgery; soft tissues measurements: PAL PD Recession (non-standardised probing). University based.
Participants Interventions
Outcomes
Notes Risk of bias Item Allocation concealment? Chung 1990 Methods Participants Interventions
Authors judgement Unclear
Description B - Unclear
RCT, split mouth, 2 treatment groups, 12 months duration. 15 individuals (8 females) commenced study and 10 completed it, aged 21-39. Test: Resorbable collagen (Perio-Barrier). Control: Modied Widman surgery. Hard tissue measurements at re-entry surgery and soft tissue measurements: PAL PD Manual probe with acrylic stent. University based and supported by Colla-Tec Corporation.
Outcomes
Notes
19
Chung 1990
(Continued)
Risk of bias Item Allocation concealment? Cortellini 1995 Methods Participants RCT, parallel design, 3 treatment groups, 12 months duration. 45 individuals, 24 females 25-61 years of age. 1 defect per individual, 15 indivduals per group. 17 incisors, 13 canines, 7 premolars & 8 molars. Test 1: Titanium reinforced non-resorbable Gore-Tex membranes removed after 6/52. Test 2: Non-absorbable, standard Gore-Tex membranes removed after 4/52. Controls: Modied Widman surgery. Maintenance intervals after initial 2 months: every month. PAL PD Recession Controlled force probing Practice based study. Authors judgement Unclear Description B - Unclear
Interventions
Outcomes
Notes Risk of bias Item Allocation concealment?
Authors judgement Yes
Description A - Adequate
Cortellini 1996 Methods Participants RCT, parallel design, 3 treatment groups, 12 months duration. 36 individuals, 23 females: 30-58 years of age. 1 infra-bony defect per individual; 12 individuals per group. Tooth population: 8 incisors, 5 canines, 8 premolars & 15 molars. Test 1: Resorbable, Resolut membrane. Test 2: Non-resobable standard Gore-Tex membrane removed after 6/52. Control: Modied Widman surgery. Maintenance intervals after initial 2 months performed every month. PAL PD Recession
20
Interventions
Outcomes
Cortellini 1996
(Continued)
Controlled force probing Notes Risk of bias Item Allocation concealment? Authors judgement Yes Description A - Adequate Practice based study.
Cortellini 1998 Methods Participants Interventions Multicentre (3 centres) RCT, split-mouth design, 2 treatment groups, 12 months duration. 23 individuals,13 females, aged 25-70. 2 defects per individual. Test: Resorbable Guidor membrane. Control: Open ap debridement. Maintenance intervals after initial 2 months performed every month. PAL PD Recession Controlled force probing Practice-based study.
Outcomes
Cortellini 2001 Methods Participants Interventions Multicentre (8 centres) RCT, parallel design, 2 treatment groups, 12 months duration. 109 individuals, age > 21 years. One defect per individual. Test: Resorbable Guideor membrane with simplied papilla preservation ap. Control: Simplied papilla preservation ap. Maintenance intervals after 3 months were every 3 months PAL PD Recession Controlled force probe
21
Outcomes
Cortellini 2001
(Continued)
Practice-based study.
Kim 1998 Methods Participants Interventions RCT, parallel design, 2 treatment groups, 12 months duration. 40 individuals aged 27-56 commenced the study and 26 (13 females) completed it. Test: Demin. bone matrix + resorbable calcium sulphate barrier. Control: Modied Widman ap. Maintenance: 3 monthly PAL PD Manual probe University based.
Outcomes
Notes Risk of bias Item Allocation concealment? Loos 2002 Methods
RCT, parallel and split mouth, 4 treatment groups, 12 months duration. All surgery performed by one operator. 25 individuals, 12 without antibiotics (mean age 26 years, 6 females), 7 smokers, 13 with antibiotics (mean age 39 years, 7 females), 5 smokers. Test: Resorbable Guidor membrane with antibiotics. Test: Guidor membrane without antibiotics. Control: Open ap debridement with antibiotics. Control: Open ap debridement without antibiotics. Maintenance intervals after 3 months, 3 monthly.
Participants
Interventions
22
Loos 2002
(Continued)
Outcomes
PAL PD Recession Controlled force probe (Florida probe) University-based.
Notes Risk of bias Item Allocation concealment? Mayeld 1998 Methods
RCT, parallel design, 2 treatment groups, 12 months duration. Surgery performed by 3 different therapists. 40 individuals commenced study and 38 (22 females) completed it. Mean ages: Test, 50.8 yr, Control, 48.7 yr. Test: Resorbable Guidor membrane. Control: Modied Widman surgery. Maintenance after initial 3 month period performed every month up to 6 months, then every 4-6 weeks. PAL PD Manual probe with acrylic stent. University based. Supported in part by Guidor AB, Huddinge, Sweden.
Participants
Interventions
Outcomes
Notes Risk of bias Item Allocation concealment? Mora 1996 Methods Participants Interventions
RCT, split-mouth design, 2 treatment groups, 12 months duration. 10 individuals, 6 females, 23-61 years of age, 2 defects per individual . Test: Non-resorbable ePTFE membranes removed after 4/52 (test). Control: Open ap debridement . Maintenance intervals after initial 3 months: every 3-4 weeks up to 6 months, then every 2-3 months.
23
Mora 1996
(Continued)
Outcomes
Hard tissue measurements at re-entry surgery. Soft tissue measurements: PAL PD Recession Manual probe, acrylic stent. University based.
Notes Risk of bias Item Allocation concealment? Pontoriero 1999 Methods Participants Interventions
RCT, split-mouth design, 4 treatment groups, 12 months duration. 40 individuals, 25 females, 32-61 years of age. Test: Absorbable Guidor. Test: Absorbable Resolut. Test: Non-absorbable Gore-Tex. Emdogain. All groups with own control (Split mouth). Maintenance at 2 weekly intervals. PAL PD Recession Controlled force probe. Practice based.
Outcomes
Notes Risk of bias Item Allocation concealment? Pritlove-Carson 1995 Methods Participants
Description D - Not used
RCT, split-mouth study, 2 treatment groups, 12 months duration. 9 individuals, 7 females, 23-47 years of age contributing 20 pairs of proximal defects.
24
Pritlove-Carson 1995
(Continued)
Interventions
Test: Non-resorbable Gore-Tex membranes removed after 5-6/52. Control: Open ap debridement. Maintenance intervals after initial 3 months: 3-4 weeks up to 6 months, then 2-3 months up to 12 months. PAL PD Recession Both manual and controlled force probing. University based. Supported by Medical Charity.
Outcomes
Ratka-Kruger 2000 Methods Participants Interventions RCT with both split-mouth and parallel group elements, 12 months. 16 individuals, 8 females, 29-61 years of age. Test: Resorbable Guidor membrane. Control: Open ap debridement. Maintenance visits at 3, 6 and 12 months. PAL PD Recession Manual probe University based.
Outcomes
Notes Risk of bias Item Allocation concealment? Sculean 2001 Methods Participants
RCT parallel design, 4 treatment groups, 12 months duration. 56 individuals, 32 females, 29-68 years of age
25
Sculean 2001
(Continued)
Interventions
Test: Resorbable Resolut membrane. Test: Emdogain only. Test: Emdogain in combination with absorbable membrane Resolut. Control: Open ap debridement. Maintenance every month after rst 3 months. PAL PD Recession Manual probe University based.
Outcomes
Notes Risk of bias Item Allocation concealment? Silvestri 2000 Methods Participants Interventions
RCT parallel design, 3 treatment groups, 12 months duration. 30 individuals, 19 females, age >/= 21 years of age. Test: Emdogain. Control: Non-resorbable Titanium reinforced ePTFE membrane. Control: Open ap debridement. Maintenance at 3 month intervals. PAL PD Recession Pressure sensitive probe Uncertain whether practice or University based.
Outcomes
26
Tonetti 1998 Methods Participants RCT multicentre (11 centres), parallel group, 2 treatment groups, 12 months duration. 143 individuals, 81 females, 26-81 years of age commenced study aged 26-81. 136 individuals completed study. Test: Resorbable Resolut membrane. Control: Open ap debridement. Maintenance every 3 months after initial 3 months. PAL PD Recession Controlled force probing Practice based. Supported in part by industry.
Interventions
Outcomes
Zucchelli 2002 Methods Participants Interventions RCT, parallel design, 3 treatment groups, 12 months duration. 90 individuals, 49 females, 30-61 years of age. Test: Emdogain, simplied papilla preservation ap (SPPF) Control: Non-resorbable Titanium reinforced ePTFE membrane with SPPF. Control: Open ap debridement SPPF .Maintenance at 1 month recall intervals. PAL PD Recession Controlled force probe Practice and University based.
Outcomes
27
Characteristics of excluded studies [ordered by study ID]
Becker 1996 Bratthall 1998 Chaves 1996 Eger 1998 Eickholz 1996 Eickholz 1998 Iversen 1996 Kilic 1997 Kim 1996 Kwan 1998 Nygaard-Ostby 1996 Quteish 1992 Shamiri 1992 Yoshinari 1996 Zybutz 2000
No control group Quasi-randomised 6 month data only No open ap debridement control 6 month data only Allocation to test and control groups was not randomised Bone density measurement technique does not correlate with radiographic assessments in other studies 6 month data only 6 month data only 6 month data only 6 month data only 6 month data only Did not employ GTR Not a randomised controlled trial Control group not randomly selected
28
DATA AND ANALYSES
Comparison 1. GTR versus control
Outcome or subgroup title 1 Attachment gain (change) 1.1 Parallel group studies 1.2 Split-mouth studies 2 Sites gaining less than 2 mm attachment 3 Probing pocket depth (change) 3.1 Parallel group studies 3.2 Split-mouth studies 4 Recession (change from baseline) 4.1 Parallel group studies 4.2 Split-mouth studies
No. of studies 16 8 8 6 11 5 6 9 4 5
No. of participants 750 472 278 424 473 255 218 407 235 172
Statistical method mean difference (Random, 95% CI) mean difference (Random, 95% CI) mean difference (Random, 95% CI) Risk Ratio (M-H, Random, 95% CI) mean difference (Random, 95% CI) mean difference (Random, 95% CI) mean difference (Random, 95% CI) mean difference (Random, 95% CI) mean difference (Random, 95% CI) mean difference (Random, 95% CI)
Effect size 1.22 [0.80, 1.64] 1.71 [1.02, 2.40] 0.79 [0.37, 1.21] 0.54 [0.31, 0.96] 1.21 [0.53, 1.88] 1.59 [0.21, 2.97] 0.87 [0.38, 1.36] 0.26 [0.08, 0.43] 0.15 [-0.12, 0.42] 0.34 [0.10, 0.57]
Analysis 1.1. Comparison 1 GTR versus control, Outcome 1 Attachment gain (change).
Review: Guided tissue regeneration for periodontal infra-bony defects
Comparison: 1 GTR versus control Outcome: 1 Attachment gain (change)
Study or subgroup
Treatment N
Control N
mean difference (SE)
mean difference IV,Random,95% CI
Weight
1 Parallel group studies Cortellini 1995 Cortellini 1996 Cortellini 2001 Mayeld 1998 Sculean 2001 Silvestri 2000 Tonetti 1998 Zucchelli 2002 30 24 55 20 14 10 69 30 15 12 54 18 14 10 67 30 2.2 (0.648) 2.6 (0.411) 0.9 (0.38) 0.2 (0.638) 1.4 (0.57) 3.6 (0.74) 0.86 (0.27) 2.3 (0.33) 4.9 % 6.6 % 6.8 % 4.9 % 5.4 % 4.3 % 7.6 % 7.2 % 2.20 [ 0.93, 3.47 ] 2.60 [ 1.79, 3.41 ] 0.90 [ 0.16, 1.64 ] 0.20 [ -1.05, 1.45 ] 1.40 [ 0.28, 2.52 ] 3.60 [ 2.15, 5.05 ] 0.86 [ 0.33, 1.39 ] 2.30 [ 1.65, 2.95 ]
Subtotal (95% CI)

-4 -2 0 2 4
47.8 %
1.71 [ 1.02, 2.40 ]
Favours control
Favours GTR
(Continued . . . )
Study or subgroup
Treatment N
Control N
Weight
(. . . Continued) mean difference

IV,Random,95% CI
Heterogeneity: Tau2 = 0.74; Chi2 = 34.37, df = 7 (P = 0.00001); I2 =80% Test for overall effect: Z = 4.87 (P < 0.00001) 2 Split-mouth studies Blumenthal 1990 Chung 1990 Cortellini 1998 Loos 2002 Mora 1996 Pontoriero 1999 Pritlove-Carson 1995 Ratka-Kruger 2000 10 10 23 25 10 30 20 11 10 10 23 25 10 30 20 11 0.42 (0.24) 1.27 (0.27) 1.4 (0.48) 0.11 (0.34) 1.3 (0.36) 1.3 (0.38) 0.02 (0.44) 0.18 (0.92) 7.8 % 7.6 % 6.1 % 7.1 % 7.0 % 6.8 % 6.4 % 3.4 % 0.42 [ -0.05, 0.89 ] 1.27 [ 0.74, 1.80 ] 1.40 [ 0.46, 2.34 ] 0.11 [ -0.56, 0.78 ] 1.30 [ 0.59, 2.01 ] 1.30 [ 0.56, 2.04 ] 0.02 [ -0.84, 0.88 ] 0.18 [ -1.62, 1.98 ]
Subtotal (95% CI)

Heterogeneity: Tau2 = 0.21; Chi2 = 18.46, df = 7 (P = 0.01); I2 =62% Test for overall effect: Z = 3.69 (P = 0.00022)
52.2 %
0.79 [ 0.37, 1.21 ]
Total (95% CI)

Heterogeneity: Tau2 = 0.54; Chi2 = 69.07, df = 15 (P<0.00001); I2 =78% Test for overall effect: Z = 5.66 (P < 0.00001)
100.0 %
1.22 [ 0.80, 1.64 ]
-4
-2
Favours control
Favours GTR
30
Analysis 1.2. Comparison 1 GTR versus control, Outcome 2 Sites gaining less than 2 mm attachment.
Comparison: 1 GTR versus control Outcome: 2 Sites gaining less than 2 mm attachment
Study or subgroup
Treatment n/N
Control n/N 2/15 2/12 17/54 11/20 22/67 7/30
Risk Ratio M-H,Random,95% CI
Weight
Risk Ratio M-H,Random,95% CI
Cortellini 1995 Cortellini 1996 Cortellini 2001 Mayeld 1998 Tonetti 1998 Zucchelli 2002
1/30 0/24 10/55 10/18 11/69 0/30
5.3 % 3.4 % 27.5 % 31.2 % 28.9 % 3.7 %
0.25 [ 0.02, 2.54 ] 0.10 [ 0.01, 2.01 ] 0.58 [ 0.29, 1.15 ] 1.01 [ 0.57, 1.79 ] 0.49 [ 0.26, 0.92 ] 0.07 [ 0.00, 1.12 ]
Total (95% CI)
226
198
100.0 %
0.54 [ 0.31, 0.96 ]
Total events: 32 (Treatment), 61 (Control) Heterogeneity: Tau2 = 0.18; Chi2 = 8.91, df = 5 (P = 0.11); I2 =44% Test for overall effect: Z = 2.10 (P = 0.036)
0.1 0.2
0.5
10
Favours GTR
Favours control
31
Analysis 1.3. Comparison 1 GTR versus control, Outcome 3 Probing pocket depth (change).
Comparison: 1 GTR versus control Outcome: 3 Probing pocket depth (change)
Study or subgroup
Treatment N
Control N
Weight
1 Parallel group studies Cortellini 2001 Mayeld 1998 Sculean 2001 Silvestri 2000 Zucchelli 2002 55 20 14 10 30 54 18 14 10 30 0.8 (0.44) 0.1 (0.62) 0.5 (0.53) 4.5 (0.54) 2 (0.34) 9.6 % 8.3 % 9.0 % 8.9 % 10.3 % 0.80 [ -0.06, 1.66 ] 0.10 [ -1.12, 1.32 ] 0.50 [ -0.54, 1.54 ] 4.50 [ 3.44, 5.56 ] 2.00 [ 1.33, 2.67 ]
Subtotal (95% CI)

Heterogeneity: Tau2 = 2.23; Chi2 = 43.60, df = 4 (P<0.00001); I2 =91% Test for overall effect: Z = 2.26 (P = 0.024) 2 Split-mouth studies Blumenthal 1990 Cortellini 1998 Loos 2002 Mora 1996 Pontoriero 1999 Ratka-Kruger 2000 10 23 25 10 30 11 10 23 25 10 30 11 0.48 (0.26) 1.3 (0.54) 0.14 (0.5) 1.8 (0.52) 1.13 (0.34) 0.23 (1.03)
46.2 %
1.59 [ 0.21, 2.97 ]
10.7 % 8.9 % 9.2 % 9.1 % 10.3 % 5.6 %
0.48 [ -0.03, 0.99 ] 1.30 [ 0.24, 2.36 ] 0.14 [ -0.84, 1.12 ] 1.80 [ 0.78, 2.82 ] 1.13 [ 0.46, 1.80 ] 0.23 [ -1.79, 2.25 ]
Subtotal (95% CI)

Heterogeneity: Tau2 = 0.16; Chi2 = 9.06, df = 5 (P = 0.11); I2 =45% Test for overall effect: Z = 3.49 (P = 0.00048)
53.8 %
0.87 [ 0.38, 1.36 ]
Total (95% CI)

Heterogeneity: Tau2 = 1.04; Chi2 = 62.90, df = 10 (P<0.00001); I2 =84% Test for overall effect: Z = 3.51 (P = 0.00044)
100.0 %
1.21 [ 0.53, 1.88 ]
-4
-2
Favours control
Favours GTR
32
Analysis 1.4. Comparison 1 GTR versus control, Outcome 4 Recession (change from baseline).
Comparison: 1 GTR versus control Outcome: 4 Recession (change from baseline)
Study or subgroup
Treatment N
Control N
Weight
1 Parallel group studies Cortellini 2001 Mayeld 1998 Sculean 2001 Zucchelli 2002 55 20 14 30 54 18 14 30 0 (0.23) 0.2 (0.292) 0 (0.491) 0.3 (0.232) 15.3 % 9.5 % 3.4 % 15.1 % 0.0 [ -0.45, 0.45 ] 0.20 [ -0.37, 0.77 ] 0.0 [ -0.96, 0.96 ] 0.30 [ -0.15, 0.75 ]
Subtotal (95% CI)

Heterogeneity: Tau2 = 0.0; Chi2 = 0.97, df = 3 (P = 0.81); I2 =0.0% Test for overall effect: Z = 1.08 (P = 0.28) 2 Split-mouth studies Blumenthal 1990 Loos 2002 Mora 1996 Pontoriero 1999 Ratka-Kruger 2000 10 25 10 30 11 10 25 10 30 11 0.28 (0.2) 0.27 (0.56) 0.35 (0.24) 0.47 (0.23) 0.14 (0.43)
43.3 %
0.15 [ -0.12, 0.42 ]
20.3 % 2.6 % 14.1 % 15.3 % 4.4 %
0.28 [ -0.11, 0.67 ] 0.27 [ -0.83, 1.37 ] 0.35 [ -0.12, 0.82 ] 0.47 [ 0.02, 0.92 ] 0.14 [ -0.70, 0.98 ]
Subtotal (95% CI)

Heterogeneity: Tau2 = 0.0; Chi2 = 0.64, df = 4 (P = 0.96); I2 =0.0% Test for overall effect: Z = 2.82 (P = 0.0048)
56.7 %
0.34 [ 0.10, 0.57 ]
Total (95% CI)

Heterogeneity: Tau2 = 0.0; Chi2 = 2.69, df = 8 (P = 0.95); I2 =0.0% Test for overall effect: Z = 2.84 (P = 0.0045)
100.0 %
0.26 [ 0.08, 0.43 ]
-1
-0.5
0.5
Favours GTR
Favours control
WHATS NEW
Last assessed as up-to-date: 12 January 2006.
13 August 2008
Amended
Converted to new review format.
33
HISTORY
Protocol rst published: Issue 1, 1999 Review rst published: Issue 2, 2001
13 January 2006 13 January 2006
New search has been performed
Searches updated.
New citation required but conclusions have not changed Six new studies have been added to the previous version of this review. The overall results however are similar and the conclusions remain unchanged. A new table of tooth loss, adverse effects and patient experience has been added.
CONTRIBUTIONS OF AUTHORS
Ian Needleman (IN), Elaine Giedrys-Leeper (EG) and Richard Tucker (RT) wrote the protocol. They independently and in duplicate assessed the eligibility of studies, extracted data and assessed the quality of the studies. Helen Worthington (HW) conducted the statistical analysis. All four review authors wrote the review.
DECLARATIONS OF INTEREST
None known.
SOURCES OF SUPPORT Internal sources

University of Manchester, UK. University College London, UK.
External sources
No sources of support supplied
34
INDEX TERMS Medical Subject Headings (MeSH)

Guided
Tissue Regeneration, Periodontal; Alveolar Bone Loss [etiology; surgery]; Bone Transplantation; Chronic Disease; Debridement [methods]; Periodontitis [ complications]; Randomized Controlled Trials as Topic
MeSH check words

Humans
35

Guided Infrabony Review Willey Blackwell

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Guided Infrabony Review Willey Blackwell

Diunggah oleh

Hak Cipta:

Format Tersedia

Guided tissue regeneration for periodontal infra-bony defects (Review)

Needleman I, Worthington HV, Giedrys-Leeper E, Tucker R

Guided tissue regeneration for periodontal infra-bony defects

The infra-bony defect and its treatment

Criteria for considering studies for this review

Search methods for identication of studies

Types of outcome measures

Data collection and analysis

Tooth loss Change in attachment levels Patient well-being or quality of life.

Attachment gain GTR only

Table 1. Summary of meta-analyses for different outcomes and sensitivity analyses

Gingival recession Gingival recession Gingival recession Gingival recession

Bone gain. Sur- GTR only gical re-entry

Bone gain. Sur- GTR + bone 1 gical re-entry substitutes

Risk of bias in included studies

Change in attachment level (Comparison 1, Outcomes 1.1, 1.2)

Table 2. Random-effects metaregression analysis of attachment change

Slope estimate -0.62

95% CI -1.46 to 0.22

Frequency of main- 16 tenance

Type of GTR barrier 18 comparisons

Gingival recession (Comparison 1, Outcome 1.4)

Cortellini 1998 Cortellini 2001

Table 3. Adverse events and patient reported outcomes

Pontoriero 1999 Pritlove-Carson 1995

Silvestri 2000 Tonetti 1998 Zucchelli 2002

Impact of study quality

protection from bias as well as factors affecting clinical heterogeneity.

Protection from bias

DISCUSSION Guided tissue regeneration alone

Adverse effects and patient-reported outcomes

Effect of study design

GTR plus bone grafts

AUTHORS CONCLUSIONS Implications for practice

Comparison with previous thorough reviews and meta-analyses

Implications for research

References to studies included in this review

References to studies excluded from this review

References to other published versions of this review

Characteristics of included studies [ordered by study ID]

Authors judgement Unclear

Notes Risk of bias Item Allocation concealment?

Authors judgement Yes

Notes Risk of bias Item Allocation concealment?

Authors judgement Yes

Notes Risk of bias Item Allocation concealment?

Authors judgement Yes

Notes Risk of bias Item Allocation concealment? Loos 2002 Methods

Authors judgement Unclear

PAL PD Recession Controlled force probe (Florida probe) University-based.

Notes Risk of bias Item Allocation concealment? Mayeld 1998 Methods

Authors judgement Unclear

Authors judgement Unclear

Authors judgement Unclear

Authors judgement Unclear

Description D - Not used

Notes Risk of bias Item Allocation concealment?

Authors judgement Yes

Authors judgement Unclear

Authors judgement Unclear

Description D - Not used

Notes Risk of bias Item Allocation concealment?

Authors judgement Unclear

Description D - Not used