The effect of regenerative periodontal therapy in preventing periodontal defects after the extraction of third molars

ORIGINAL CONTRIBUTIONS

The effect of regenerative periodontal therapy in preventing periodontal defects after the extraction of third molars A systematic review and meta-analysis Chun-Teh Lee, DDS, MS, DMSc; Lauren Hum, DMD; Ya-Wei Chen, DDS

C

linicians remove more than 10 million partially or fully impacted third molars every year in the United States.1,2 The indications for third-molar extraction include pain, inflammation, pathology associated with tooth follicle, nonrestorable tooth, facilitation of dental procedures, such as orthodontic treatments, and prevention of disease involvement on adjacent teeth.3,4 The results of several studies have demonstrated that retained asymptomatic impacted third molars were associated with the periodontal pathology of the second molars.5-8 The prevalence of a probing depth (PD) of at least 5 millimeters on the distal aspect of mandibular second molars was 5 times higher than in maxillary second molars when second molars had adjacent asymptomatic third molars.9 Furthermore, the PD on the distal aspect of mandibular second molars progressed more easily in mandibular second molars than in maxillary second molars.10 Routine mechanical debridement did not significantly reduce the PD at sites with a PD of 4 mm or greater between the mandibular second and third molars.11 Therefore, many mandibular third molars are extracted to treat periodontal pathology or to prevent progression of periodontal disease at the mandibular second molars. However, investigators often found a periodontal defect on the distal site of the mandibular second molar after third-molar extraction.12-15 According to the results of a retrospective study,14 more than 40% of mandibular second molars had intrabony defect of

ABSTRACT Background. Periodontal defect on the distal aspect of mandibular second molars is a common complication after mandibular third-molar extraction. Researchers have proposed different procedures, but no evidence has shown that a single effective method can prevent or treat this complication. Methods. The authors conducted a systematic review and meta-analysis to answer this clinical question: what is the effect of regenerative periodontal therapy on the periodontal tissue healing of the distal site of the mandibular second molar after impacted mandibular third-molar extraction compared with extraction alone without using any biomaterials during a follow-up period of at least 6 months? The authors conducted an electronic search for randomized controlled trials using MEDLINE, Embase, and other databases, and they assessed the quality of selected articles. Results. Among the 1,083 eligible articles found in the initial search, 7 studies fit all of the selection criteria. All of these studies had a follow-up period lasting at least 6 months. The authors found that regenerative periodontal therapy was significantly more effective in gaining clinical attachment level or reducing probing depth at the distal site of the mandibular second molar than extraction without therapy (weighted mean difference of clinical attachment level gain, 1.94 millimeters [95% confidence interval {CI}, 1.56-2.31]; weighted mean difference of probing depth reduction, 1.67 mm [95% CI, 1.15-2.19]). Conclusions and Practical Implications. The results of our systematic review and meta-analysis demonstrated that regenerative periodontal therapy effectively prevents the periodontal defect associated with impacted mandibular third-molar extraction. Clinicians should consider performing guided tissue regeneration when the defect is anticipated. Key Words. Guided tissue regeneration; periodontal disease; third molars; evidence-based dentistry; outcome assessment. JADA 2016:-(-):--http://dx.doi.org/10.1016/j.adaj.2016.03.005

Copyright ª 2016 American Dental Association. All rights reserved.

JADA

( )

- -

http://jada.ada.org

-

2016 1

ORIGINAL CONTRIBUTIONS

at least 4 mm, and more than 50% of mandibular second molars had a PD of at least 7 mm even 4 years after third-molar extraction. The questions then become whether the relatively high prevalence of having a periodontal defect on the mandibular second molar after impacted third-molar extraction outweighs the treatment effect of extraction and whether this complication is preventable. To prevent the residual periodontal defect distal to the mandibular second molar after partially or fully impacted third-molar extraction, investigators have proposed the following procedures: scaling and root planing on the distal site of second molar,16-18 using a specific flap design during extraction procedure,19,20 and performing regenerative periodontal therapy on the extraction site.21,22 Regenerative periodontal therapy attempts to restore lost periodontal structures and functional attachment through the regeneration of cementum, periodontal ligament, and alveolar bone.23 Clinicians can evaluate the outcomes by measuring changes of clinical attachment level (CAL), PD, and bony defect. Frequently conducted techniques of regenerative therapy include osseous grafting, guided tissue regeneration (GTR), and use of biologics (for example, growth factors, enamel matrix derivative, platelet-rich plasma). Space provision, wound stability, and cell induction are key factors in the periodontal regeneration that can be achieved by using these techniques.24 Given that this procedure has promising clinical outcomes in treating general periodontal defects, regenerative periodontal therapy appears to be a practical and predictable treatment to prevent the periodontal complication after third-molar extraction.25,26 Investigators of studies evaluating the effectiveness of regenerative periodontal therapy on the prevention of periodontal defects on the distal site of mandibular third molars after the extraction of third molars have reported varying results.21,22,27,28 The purpose of our study was to systematically review the literature and conduct a metaanalysis of data to assess the outcome of regenerative periodontal therapy in preventing the loss of periodontal tissue distal to mandibular second molars after the extraction of impacted mandibular third molars. METHODS

Focused question. We developed our focus question by addressing the population, the intervention (or exposure), the appropriate control group (or comparator), the outcomes of interest, and the study design. Our question was as follows: what is the effect of regenerative periodontal therapy—including osseous grafting, GTR, use of growth factors, enamel matrix derivative, or plateletrich plasma—on the periodontal tissue healing of the distal site of the mandibular second molar after impacted mandibular third-molar extraction compared with extraction alone without using any biomaterials during a follow-up period of at least 6 months?

2 JADA

( )

- -

http://jada.ada.org

-

2016

Study selection criteria. We included only articles that were published in English and described prospective randomized or nonrandomized controlled trials. Each of the included studies had at least 2 patient groups, and the number of patients in each group was no fewer than 10. All of the patients in the included studies had undergone mandibular third-molar extraction. Patients in the regenerative therapy group received the regenerative periodontal therapy in the sockets after mandibular third-molar extraction, and patients in the control group had extraction sockets that had healed naturally without receiving any biomaterials. We included studies whose investigators reported the surgical method of extraction, the change of CAL, the PD of the distal site of mandibular second molars, or a combination of these, during a follow-up period of at least 6 months. Search strategy. We conducted the search in electronic databases, including MEDLINE (PubMed), Embase, Web of Science, and Dental and Oral Sciences Source, from January 1960 to August 2015. Appendix 1 (available online at the end of this article) lists the search strategies we used in these databases. In addition to the searches in electronic databases, we searched the archives of the following journals: Journal of Periodontology, Journal of Clinical Periodontology, International Journal of Periodontics and Restorative Dentistry, Journal of Maxillofacial and Oral Surgery, International Journal of Maxillofacial and Oral Surgery, and British Journal of Oral and Maxillofacial Surgery. Moreover, we screened the reference lists of selected articles to find additional articles that might fit the selection criteria. Quality assessment. We assessed the randomized controlled trials (RCTs) using the Cochrane Collaboration’s tool.29 We assigned a category of “high risk of bias,” “low risk of bias,” or “unclear risk of bias” to each assessment item (Appendix 2, available online at the end of this article, describes the quality assessment of the RCTs). In addition, we determined the evidence level of the study by using the Oxford Centre for EvidenceBased Medicine recommendation.28-30 Two authors (C.-T.L., Y.-W.C.) independently performed the assessment, and they analyzed the interexaminer agreement by using the k statistic. The 2 examiners resolved any discrepancy in the quality assessment via discussion. Data extraction and data synthesis. After the initial search process, 2 authors (C.-T.L., Y.-W.C.)

ABBREVIATION KEY. BPBM: Bovine porous bone mineral. CAL: Clinical attachment level. DBP: Demineralized bone powder. DFDBA: Demineralized freeze-dried bone allograft. GTR: Guided tissue regeneration. LC: Lincomycin. NA: Not applicable. PD: Probing depth. RCT: Randomized controlled trial. SRP: Scaling and root planing. WDCAL: Weighted mean difference of clinical attachment level gain. WDPD: Weighted mean difference of probing depth reduction.

ORIGINAL CONTRIBUTIONS

independently screened the titles and abstracts of Articles identified after search in articles and excluded the databases and journals articles that had unre(N = 1,083) lated topics. They read full texts of the potenArticles excluded with unrelated topics or primary outcomes tially qualified articles and then included the qualified studies. They Articles identified by authors after resolved disagreements they reviewed titles and abstracts by discussion. In situa(n = 30) tions for which more Eligibility criteria than 1 article had the • Published in English • Randomized controlled trial same patient cohort, • The studies had at least 2 patient they included only the groups, and the number of patients in article that had the each group was no fewer than 10. longest follow-up period All of the patients had undergone or the largest sample mandibular third-molar extraction. size. • The patients in the regenerative Two authors (C.-T.L., therapy group received the regenerative Y.-W.C.) independently peridontal therapy in the sockets after mandibular third-molar extraction, and extracted data by using a the patients in the control group had specially designed data extraction sockets that had healed extraction form. Another naturally without receiving any author (L.H.) confirmed biomaterials. the accuracy of extracted • The studies whose investigators data. We contacted the reported the surgical method of authors of the included extraction, the change of clinical attachment level, the probing depth of articles if we felt that inthe distal site of the mandibular second formation was unclear molar, or a combination of these during and had to be clarified. a follow-up period of at least 6 months. Data analysis. The primary outcome was the difference of CAL gain at the distal site Articles excluded because of having Articles eligible for data analysis of mandibular second the same patient cohort (n = 1) (n = 8) molars after impacted mandibular third-molar extraction between the regenerative therapy group and the control Articles included in the final data group. We defined CAL analysis (n = 7) as the distance from the cementoenamel junction of the tooth to the tip Figure 1. Search strategy and screening process. of a periodontal probe during diagnostic probing. We calculated CAL gain by subtracting the CAL positive value indicated that the regenerative therapy at the last measurement from the CAL at baseline. We group had more CAL gain than the control group and calculated the weighted mean difference of CAL gain vice versa. The secondary outcome was the difference of (WDCAL) by subtracting the CAL gain at the distal site PD reduction between the 2 groups. We defined PD as of the mandibular second molar in the control group the distance from the gingival margin to the tip of the from the CAL gain at the distal site of mandibular sec- periodontal probe during diagnostic probing. We calcuond molars in the regenerative therapy group. Then we lated the PD reduction by subtracting the PD at the last pooled the differences of CAL gain between 2 groups measurement from the PD at baseline. We calculated the in each study on the basis of the assigned weights. A weighted mean difference of PD reduction (WDPD) by

JADA

( )

- -

http://jada.ada.org

-

2016 3

ORIGINAL CONTRIBUTIONS

TABLE 1

Summary of selected studies. STUDY

NO. OF PATIENTS WHO DROPPED OUT OR WERE LOST DURING FOLLOW-UP*

AGE (MEAN), Y

OSTECTOMY

Unclear, horizontal position

2

26-50

Yes

Soft-tissue impacted, not mentioned

6

> 25 (31)

Yes

Treat group 1 (DBP¶¶): 12, 12 Not mentioned, not mentioned Both, various Control group 1***: 12, 12 positions Control group 2***: 12, 12

3

> 26.0 (29.7)

Yes

Treatment group: 15, 15

Soft-tissue impacted, mesioangular position

0

(24.9)†††

Yes

Both, mesioangular or horizontal position

0

21-30

Yes

Unclear, horizontal position

0

30-35 (32)

Yes

Unclear, mesioangular, or horizontal position

0

25-30 (26.5)

Yes

STUDY DESIGN

Pecora and RCT‡‡ Colleagues,21 1993

NO. OF PATIENTS, NO. OF TEETH*

20 patients, 20 total teeth Treatment group: 10, 10

CHARACTERISTICS OF IMPACTED MANDIBULAR THIRD MOLAR AND ANGULATION†

Control group: 10, 10 Throndson and Sexton,35 2002

RCT, split-mouth

Dodson,36 2004

RCT, split-mouth

Treatment group: 14, 14 Control group: 14, 14

Treat group 2 (GTR##): 12, 12

Aimetti and RCT, split-mouth Colleagues,22 2007 Sammartino and RCT Colleagues,37 2009

Control group: 15, 15 45 patients, 90 total teeth Treat group 1 (BPBM‡‡‡ alone): 30, 30 Treatment group 2 (BPBM þ collagen membrane): 30, 30 Control group: 30, 30

Hassan and RCT, split-mouth Colleagues,38 2012

Control group: 14, 14

Treatment group: 14, 14

Tabrizi and RCT split-mouth Colleagues,39 2014

Treatment group 1 (DFDBA§§§): 10, 10 Treatment group 2 (DFDBA þ LC¶¶¶): 10, 10 Control group 1***: 10, 10 Control group 2***: 10, 10

* The reported number of patients did not include the number of patients who dropped out or were lost during follow-up. † All the mandibular third molars in the included studies were impacted. The impaction status could be categorized as soft-tissue impacted, bony impacted, both (soft-tissue impacted and bony impacted), or unclear if the impaction status was not clearly mentioned. ‡ SRP: Scaling and root planing. § All the time points of clinical measurements were listed. ¶ PD: Probing depth. # PD reduction was defined as subtracting the PD at the last measurement from PD at baseline. ** CAL: Clinical attachment level. †† CAL gain was defined by subtracting the CAL at the last measurement from the CAL at baseline. ‡‡ RCT: Randomized controlled trial. §§ NA: Not applicable. ¶¶ DBP: Demineralized bone powder. ## GTR: Guided tissue regeneration. *** The control group 1 or the control group 2 matched the test group 1 or the test group 2. ††† Only mean age is available. ‡‡‡ BPBM: Bovine porous bone mineral. §§§ DFDBA: Demineralized freeze-dried bone allograft. ¶¶¶ LC: Lincomycin.

subtracting the PD reduction at the distal site of the mandibular second molar in the control group from the PD reduction at the distal site of the mandibular second molar in the regenerative therapy group. Then we pooled the differences of PD reduction between the 2 groups in each study on the basis of the assigned

4 JADA

( )

- -

http://jada.ada.org

-

2016

weights. A positive value indicated that the regenerative therapy group had more PD reduction than the control group and vice versa. If the study included multiple groups that qualified for data analysis, we extracted the data from each group and treated each data set independently.

ORIGINAL CONTRIBUTIONS

TABLE 1 (CONTINUED)

SOCKET DEBRIDEMENT AND SRP‡ AT DISTAL SURFACE OF SECOND MOLAR

BONE GRAFTS

MEMBRANE

FOLLOWUP PERIOD (MO)§

MEAN (STANDARD DEVIATION) PD¶ REDUCTION IN THE TREATMENT GROUP (MILLIMETERS)#

3, 6, 9, 12 5.7 (2.0) Expanded polytetrafluoroethylene membranes

MEAN (STANDARD DEVIATION) PD REDUCTION IN THE CONTROL GROUP (mm)#

MEAN (STANDARD DEVIATION) CAL** GAIN IN THE TREATMENT GROUP (mm)††

MEAN (STANDARD DEVIATION) CAL GAIN IN THE CONTROL GROUP (mm)††

4.00 (1.69)

4.30 (2.45)

1.90 (1.73)

1.11 (1.30)

0.25 (0.85)

Yes, yes

None

Yes, unclear

None Bioactive glass particles (BioGran, Orthovita)

3, 6, 12

Yes, yes

Allograft (DPB) (test group 1)

Test 1: 3.6 (3.6) 0.25, 0.75, 3.00, 6.50 Test 2: 1.9 (3.1)

Control 1: 3.3 (2.1)

Test 1: 3.7 (3.5)

Control 1: 3.1 (2.1)

Control 2: 1.9 (2.5)

Test 2: 1.8 (2.9)

Control 2: 1.2 (2.4)

0.54 (0.31)

2.88 (0.44)

0.52 (0.38)

Test 1: 2.65 (0.58)

0.74 (0.46)

Collagen membrane (test group 2)

NA§§

2.9 (0.5)

NA

Yes, yes

None

Polyglycolic acid, polylactic acid

12

Yes, yes

Xenograft (BPBM)

Collagen membrane

3, 6, 9, 12, Test 1: 2.28 (0.69) 18, 24, 36, 48, Test 2: 3.25 (0.58) 60, 72

0.73 (0.43)

Test 2: 3.20 (0.57)

Yes, yes

Xenograft (BPBM)

Collagen membrane

3, 6, 9, 12 4.40 (0.61)

2.9 (0.7)

Unclear, unclear

Allograft (DFDBA)

None

6.5

Test 1: 0.9 (1.66)

Control 1: 1.35 (0.43)

Test 2: 1.05 (1.61)

Control 2: 1.34 (0.42)

The investigators of some studies did not provide standard deviations (SD) of the difference of CAL gain or PD reduction between the regenerative therapy group and the control group. We imputed the missing SDs by using an assumptive correlation coefficient that equaled 0.5, or the known t value.31-33 Owing to the heterogeneity of the data sets in the primary and secondary outcomes (c2 ¼ 25.41, P < .01, and c2¼ 55.52, P < .01, respectively), we chose to use the random-effects model (that is, the DerSimonian and Laird test) for our meta-analysis.34 We also assessed heterogeneity between data sets by using I2 statistics describing the variation of each data set. We generated forest plots to demonstrate the individual and pooled-effect estimates, as well as 95% confidence intervals (CI). We performed meta-regression to evaluate the correlation between the outcome of interest and different variables. To evaluate the variables causing heterogeneity or result bias, we conducted subgroup analysis. We performed sensitivity analysis

3.01 (1.03)

1.25 (0.65)

NA

NA

to assess the robustness of the results from metaanalysis. We evaluated publication bias by using the Egger test and funnel plot. We performed all statistical analysis using Stata (Version 11.2, StataCorp). We defined statistical significance as P < .05. RESULTS

Study selection. We found 1,083 articles by following the search process (Figure 1). After reviewing the titles and abstracts, we identified 30 articles and excluded 23 articles from the final analysis because of 1 or more of the following reasons: incomplete data for outcomes of interest, no control group, a follow-up period shorter than 6 months, more than 1 study had the same cohort, and the study had an insufficient number of patients (Appendix 3, available online at the end of this article, lists the excluded studies). We included 7 articles for data analysis21,22,35-39; however, we selected only 6 studies21,22,35-38 to calculate WDCAL and only 6 studies21,22,36-39 to calculate WDPD.

JADA

( )

- -

http://jada.ada.org

-

2016 5

ORIGINAL CONTRIBUTIONS

Study

Mean Difference (95% CI)

Weight, %

2.40 (0.54-4.26)

3.48

0.60 (–1.53 to 2.73)

2.74

Guided Tissue Regeneration Pecora and Colleagues,21 1993 36

Dodson,

2004 22

Aimetti and Colleagues,

2007

2.36 (2.07-2.65)

21.49

2.19 (1.47-2.91)

27.70

Throndson and Sexton,35 2002

0.86 (0.05-1.67)

11.40

Dodson,36 2004

0.60 (–1.71 to 2.91)

2

Subtotal: P = .275; I = 22.5% Osseous Grafting

2.37

1.91 (1.65-2.17)

22.04

1.36 (0.45-2.27)

35.81

Sammartino and Colleagues,37 2009

2.46 (2.20-2.72)

22.09

Hassan and Colleagues,38 2012

1.76 (1.12-2.40)

14.39

Subtotal: P = .047; I = 74.7%

2.17 (1.50-2.85)

36.49

Overall: P = .001; I2 = 72.5%

1.94 (1.56-2.31)

100.00

Sammartino and Colleagues,37 2009 2

Subtotal: P = .033; I = 70.7% Combination

2

–2

0

2

4

5

Figure 2. Weighted mean difference of clinical attachment level gain on the distal site of mandibular second molars (guided tissue regeneration, osseous grafting, and combination). Weights are from random-effects model analysis. CI: Confidence interval.

We excluded 1 study39 from WDCAL analysis, and we excluded 1 study35 from WDPD analysis because the investigators of these 2 studies35,39 did not provide the clinical information we required. The procedures used in these studies included osseous grafting, GTR, and combination technique (Table 121,22,35-39). We defined the combination technique as placing grafts in the defect and covering the defect with a barrier at the same time. None of the included studies used growth factors, enamel matrix derivative, or platelet-rich plasma in regenerative periodontal therapy. Quality assessment and heterogeneity evaluation. In the assessment we conducted using the Cochrane Collaboration’s tool, we found that the investigators of these studies determined that 3 to 7 of 9 assessment items had a low risk of bias (eTable,21,22,35-39 available online at the end of this article). According to Oxford Centre for Evidence-Based Medicine30 recommendations, the evidence level of each study was generally 2b, but 1 study had 1b evidence level.37 For these 2 assessments, the k statistics between 2 of the authors (C.-T.L., Y.-W.C.) were 0.94 and 1.0, respectively. Those data sets for WDCAL and WDPD analyses had considerable heterogeneity (I2 ¼ 72.5, P < .01; and I2 ¼ 85.6, P < .01).40

6 JADA

( )

- -

http://jada.ada.org

-

2016

Study description. Five studies were RCTs with splitmouth design.22,35,36,38,39 In these studies, the patients had bilateral mandibular third molars that were randomly assigned to receive 1 type of therapy. Two studies were RCTs without the split-mouth design.21,37 In these studies, 1 mandibular third molar in each patient was randomly assigned to a specific group. Six of 7 studies reported the position of impacted mandibular third molars: mesioangular22; horizontal21,38; mesioangular and horizontal37,39; and mesioangular, horizontal, distal, and vertical.36 The investigators of 1 study35 did not specify the position of the mandibular third molars (Table 121,22,35-39). The biomaterials used in the studies were heterogeneous. Three of the included studies had multiple groups whose study participants received different types of regenerative periodontal therapy,36,37,39 and the data set was analyzed independently. In some studies, investigators used xenografts,37,38 allografts,36,39 or alloplasts35 to perform osseous grafting or a combination technique. Some study investigators used nonresorbable membrane (expanded polytetrafluoroethylene)21 or resorbable collagen membrane22,36-38 to perform GTR or a combination technique (Table 121,22,35-39).

ORIGINAL CONTRIBUTIONS

Mean Difference (95% CI) Weight, %

Study Guided Tissue Regeneration Pecora and Colleagues,21 1993 36

Dodson,

2004 22

Aimetti and Colleagues,

2007

2

Subtotal: P = .094; I = 57.8%

1.70 (0.15-3.25)

6.97

0.00 (–2.25 to 2.25)

4.10

2.36 (2.06-2.66)

17.37

1.76 (0.61-2.90)

28.43

Osseous Grafting Dodson,36 2004

0.30 (–2.06 to 2.66) 37

Sammartino and Colleagues,

3.82

1.55 (1.26-1.84)

2009

Tabrizi and Colleagues,39 2014 (demineralized freeze-dried bone allograft)

17.42

–0.70 (–2.03 to 0.63)

8.33

0.90 (–0.40 to 2.20)

8.53

0.65 (–0.50 to 1.81)

38.09

Sammartino and Colleagues,37 2009

2.52 (2.26-2.78)

17.62

Hassan and Colleagues,38 2012

2.50 (2.01-2.99)

15.85

Subtotal: P = .943; I2 = 0%

2.52 (2.29-2.74)

33.47

Overall: P = 0; I2 = 85.6%

1.67 (1.15-2.19)

100.00

Tabrizi and Colleagues,39 2014 (demineralized freeze-dried bone allograft plus lincomycin) Subtotal: P = .007; I2 = 75.1% Combination

–3

–2

0

2

4

Figure 3. Weighted mean difference of probing depth reduction on the distal site of mandibular second molars (guided tissue regeneration, osseous grafting, and combination). Weights are from random-effects model analysis. CI: Confidence interval.

In our review, we found that the investigators of 4 studies reported a change in the alveolar bone level on the distal aspect of mandibular second molars between the baseline measurement and the last measurement in the regenerative therapy group and the control group.22,35,38,39 Owing to the heterogeneity of measurement methods and insufficient data, we could not systematically analyze the change in the alveolar bone level. Synthesis of results. We calculated the CAL gain and PD reduction on the distal aspect of mandibular second molars after mandibular third-molar extraction. Regenerative periodontal therapy was significantly more effective in gaining CAL at the distal site of mandibular second molars than extraction alone (WDAL, 1.94; 95% CI, 1.56-2.31) (Figure 221,22,35-38). Regenerative periodontal therapy also was significantly more effective in reducing PD at the distal site of mandibular second molars than extraction alone (WDPD, 1.67; 95% CI, 1.15-2.19) (Figure 321,22,36-39).

Meta-regression analysis. We evaluated several clinical variables (for example, types of regenerative periodontal therapy, follow-up period, mean CAL at baseline, mean PD at baseline) in the meta-regression analysis. Osseous grafting was significantly associated with the difference of PD reduction between 2 groups (P ¼ .04). The results showed that osseous grafting had the least effect on PD reduction in the regenerative periodontal group compared with GTR and the combination technique. Other variables were not related significantly to the difference of PD reduction and the difference of CAL gain between 2 groups (Table 2). Subgroup analysis and sensitivity analysis. The type of regenerative procedures affected the clinical outcomes on the basis of the results of our meta-regression analysis. According to different regenerative procedures conducted in the included studies, we categorized these data sets into the osseous grafting group,35-37,39 the GTR group,21,22,36 and the combination group in the subgroup analysis.37,38

JADA

( )

- -

http://jada.ada.org

-

2016 7

ORIGINAL CONTRIBUTIONS

TABLE 2

barrier membranes increased more CALs and reduced more PDs on the intrabony defects VARIABLES WEIGHTED MEAN DIFFERENCE WEIGHTED MEAN DIFFERENCE than grafting alone.41 HowevOF CLINICAL ATTACHMENT OF PROBING DEPTH er, in the histologic results, the LEVEL GAIN REDUCTION combination technique did Coefficient 95% CI* P Coefficient 95% CI P not demonstrate a better Value Value regeneration ability than 0.70 1.72 to 0.31 .14 1.42 2.76 to 0.07 .04 Osseous Grafting† osseous grafting or GTR 0.01 0.01 to 0.03 .31 0.02 0.02 to 0.05 .31 Follow-up Period‡ alone.26 In general, we found ¶ Baseline Clinical 0.07 0.98 to 0.84 .86 NA NA NA § that none of the techniques— Attachment Level that is, osseous grafting, GTR, Baseline Probing Depth NA NA NA 0.30 0.14 to 0.73 .15 or a combination technique— * CI: Confidence interval. † The positive value of coefficient means that the technique of osseous grafting was related positively to the was significantly superior to difference of clinical attachment level gain (probing depth reduction) between 2 groups (the clinical the others.42 In the subgroup attachment level gain and probing depth reduction was more favorable in the regenerative therapy group than in the control group). analysis and meta-regression ‡ The positive value of the coefficient means that the length of the follow-up period was related positively analysis of this review, we to the difference of clinical attachment level gain (probing depth reduction) between 2 groups. noted that osseous grafting § The positive value of the coefficient means that the mean baseline clinical attachment level (mean baseline probing depth) of 2 groups was related positively to the difference of clinical attachment level had inferior outcomes in CAL gain (probing depth reduction) between 2 groups. gain and PD reduction when ¶ NA: Not applicable. compared with the other techniques. Use of a memThe osseous grafting group had the lowest WDCAL brane appeared to be critical in improving the periodontal (1.36; 95% CI, 0.45-2.27) and WDPD (0.65; 95% CI, 0.50 healing of the defect adjacent to mandibular second molar to 1.81) compared with the GTR group (WDCAL, 2.19; after the impacted third-molar extraction. However, some 95% CI, 1.47-2.91; and WDPD, 1.76; 95% CI, 0.61-2.90) factors potentially biased the results. For example, in the and the combination group (WDCAL, 2.17; 95% CI, 1.50- analysis of CAL change, we included 1 study whose in2.85; and WDPD, 2.52; 95% CI, 2.29-2.74) (Figures 221,22,35-38 vestigators used bioglass.35 Clinicians usually have and 321,22,36-39). considered bioglass and other alloplasts as inferior bioIn the sensitivity analysis of assessing the impact of materials for periodontal regeneration when compared each data set, we found that WDCAL was the lowest with other materials.23,26 Also, in the analysis of PD when 1 study22 was omitted (1.78; 95% CI, 1.30-2.26). reduction, we included 1 study39 whose investigators 35 WDCAL was the highest when 1 study was omitted noted a minor periodontal defect at baseline and whose (2.11; 95% CI, 1.80-2.43). WDPD was the lowest when results demonstrated less favorable clinical outcomes than 1 study22 was omitted (1.44; 95% CI, 0.79-2.10). WDPD the others. It is known that the clinical benefits of was the highest when the data set of the demineralized regenerative therapy are limited when the defects are 39 freeze-dried bone allograft group from the other study shallow (# 4 mm).43 The use of an alloplast and the was omitted (1.94; 95% CI, 1.48-2.41). No individual data presence of a shallow defect at baseline in the included set appeared to significantly affect the outcomes. studies might affect the clinical outcomes in the osseous Publication bias. The results of the Egger test showed grafting group. Owing to the limited number of the no significance in regards to WDCAL and WDPD included studies and potential bias, we do not recommend (P ¼ .14 and P ¼ .11, respectively), but the funnel plots any specific regenerative therapy to treat periodontal did not demonstrate symmetric distribution of the data defect after impacted mandibular third-molar extraction. sets (eFigure 1 and eFigure 2, available online at the end It is worth noting that, except for 2 studies,35,39 the of this article). Owing to the limited number of included study participants in the included studies had initial CAL studies, we could not completely rule out the possibility or PD of at least 5 mm. Although the mean baseline CAL of publication bias. or mean baseline PD were not related significantly to WDCAL or WDPD, the favorable clinical effect of DISCUSSION regenerative periodontal therapy might not have been revealed when the distal defects of mandibular second The results of our review indicated that regenerative periodontal therapy is more effective in treating the distal molars were shallow before impacted mandibular thirdperiodontal defect of mandibular second molars after molar extraction. In other words, if there was no deep mandibular third-molar extraction than extraction alone. PD or significant attachment loss between the impacted mandibular third molar and the second molar, it might In our review of articles regarding regenerative periodontal therapy in treating general periodontal defects, we not have been necessary to perform regenerative therapy after extraction. found that the use of bone grafts in combination with

Meta-regression analysis.

8 JADA

( )

- -

http://jada.ada.org

-

2016

ORIGINAL CONTRIBUTIONS

Even with the promising results of CAL gain and PD reduction, it is difficult to conclude that regeneration of periodontal tissue happens at the distal site of mandibular second molars without seeing histologic results. The investigators of only 1 of the included studies37 descriptively reported the histologic results of mandibular third-molar extraction sites. The results indicated that residual xenogenic grafts were present in all bone cores harvested 6 months after extraction and that combination therapy appeared to have stronger osteoblastic activity as well as fewer residual grafts than grafting alone. More histologic results are expected in the future to prove the regeneration ability of these materials on the periodontal defect at the distal site of mandibular second molars. Investigators have considered the number of bony walls,44,45 the depth of defects,43 and the angulation of defects46 to have an impact on the healing of intrabony defects treated by regenerative therapy. The abundance of cellular and vascular resources might cause the effect of bony configurations on the regenerative outcomes.26 In this review, we found that the investigators of 2 studies22,38 reported that the change of alveolar bone level on the distal aspect of mandibular second molars between the baseline measurement and the last measurement in the regenerative therapy group was more significant than the change in the control group. Conversely, the investigators of 2 other studies35,39 did not show a significant difference between groups. There was no description of bony configurations in these sites, but the initial depth of defect at the distal site of the mandibular second molar in the 2 studies with limited difference of alveolar bone level between 2 groups were shallower than the level found by the investigators of the other 2 studies.22,38 In addition, we observed the limited differences of CAL and PD between 2 groups in 2 of the studies.35,39 Owing to the heterogeneous data provided by the investigators of only 4 studies, we could make no conclusion regarding the effect of regenerative periodontal therapy on the healing of bone in this review. Investigators have demonstrated the long-term stability of clinical parameters in well-maintained teeth treated by regenerative periodontal therapy in several studies that had a follow-up period of more than 5 years.47-50 In this review, the investigators of only 1 included study37 reported follow-up clinical outcomes for longer than 1 year. The PD and CAL at the distal site of mandibular second molars in the regenerative group and the control group were stable between the first year and the sixth year. Additional studies with longer follow-up periods are needed to support the long-term stability of periodontal healing at the distal defect of mandibular second molars after regenerative periodontal therapy. Age is a clinical factor associated with the healing of defect after mandibular third-molar extraction. Patients younger than 25 to 30 years are considered to have

a threshold favoring the defect healing process.51,52 Given that most of the investigators of the included studies recruited patients older than 25 years and that the mean age of patients in these studies was similar, we did not analyze the age of patients in this review. CONCLUSION

Regenerative periodontal therapy is more effective in increasing CAL and reducing PD on the distal aspect of mandibular second molars after impacted mandibular third-molar extraction than extraction alone. Osseous grafting appears to be a less effective regenerative therapy than GTR or a combination technique. Additional RCTs are needed to investigate whether a specific biomaterial used in regenerative periodontal therapy could have better clinical outcomes in resolving periodontal defect after impacted mandibular third-molar extraction than others. n SUPPLEMENTAL DATA

Supplemental data related to this article can be found at: http://dx.doi.org/10.1016/j.adaj.2016.03.005. Dr. Lee is an assistant professor, Department of Periodontics and Dental Hygiene, University of Texas Health Science Center at Houston, Houston, TX. Dr. Hum is a dental student, Harvard School of Dental Medicine, Boston, MA. Dr. Chen is an attending surgeon, Oral and Maxillofacial Surgery, Department of Stomatology, Taipei Veterans General Hospital, and an assistant professor, School of Dentistry, National Yang-Ming University, Taipei, Taiwan. Address correspondence to Dr. Chen, Oral and Maxillofacial Surgery, Department of Stomatology, Taipei Veterans General Hospital, No. 201, Sec. 2, Shipai Road, Beitou District, Taipei, Taiwan 112, e-mail [email protected]. Disclosure. None of the authors reported any disclosures. The authors express sincere appreciation to Dr. Paul Bain, reference and education services librarian, The Francis A. Countway Library of Medicine (an alliance of the Boston Medical Library and Harvard Medical School), for his valuable contribution in the search strategy of this article. 1. Friedman JW. The prophylactic extraction of third molars: a public health hazard. Am J Public Health. 2007;97(9):1554-1559. 2. American Dental Association Survey Center. 2005-2006 Survey of Dental Services Rendered: Dental Practice. Chicago, IL: American Dental Association; 2007. Available at: http://www.ada.org/w/media/ADA/ Science%20and%20Research/HPI/Files/05_sdsr20140404t121935. Accessed March 20, 2016. 3. Haug RH, Abdul-Majid J, Blakey GH, White RP. Evidenced-based decision making: the third molar. Dent Clin North Am. 2009;53(1):77-96, ix. 4. American Association of Oral and Maxillofacial Surgeons (AAOMS). White paper on third molar data. Available at: http://www.aaoms.org/ images/uploads/pdfs/white_paper_third_molar_data.pdf. Accessed March 20, 2016. 5. Elter JR, Cuomo CJ, Offenbacher S, White RP Jr. Third molars associated with periodontal pathology in the Third National Health and Nutrition Examination Survey. J Oral Maxillofac Surg. 2004;62(4): 440-445. 6. Elter JR, Offenbacher S, White RP, Beck JD. Third molars associated with periodontal pathology in older Americans. J Oral Maxillofac Surg. 2005;63(2):179-184.

JADA

( )

- -

http://jada.ada.org

-

2016 9

ORIGINAL CONTRIBUTIONS

7. Marciani RD. Is there pathology associated with asymptomatic third molars? J Oral Maxillofac Surg. 2012;70(9 suppl 1):S15-S19. 8. Nunn ME, Fish MD, Garcia RI, et al. Retained asymptomatic third molars and risk for second molar pathology. J Dent Res. 2013;92(12): 1095-1099. 9. Blakey GH, Marciani RD, Haug RH, et al. Periodontal pathology associated with asymptomatic third molars. J Oral Maxillofac Surg. 2002; 60(11):1227-1233. 10. Blakey GH, Jacks MT, Offenbacher S, et al. Progression of periodontal disease in the second/third molar region in subjects with asymptomatic third molars. J Oral Maxillofac Surg. 2006;64(2):189-193. 11. Fisher EL, Blakey GH, Offenbacher S, Phillips C, White RP Jr. Mechanical debridement of subgingival biofilm in participants with asymptomatic third molars does not reduce deeper probing depths in the molar regions of the mouth. J Oral Maxillofac Surg. 2013;71(3):467-474. 12. Ash M, Costich E, Hayward J. A study of periodontal hazards of third molars. J Periodontol. 1962;33(3):209-219. 13. Kugelberg CF, Ahlstrom U, Ericson S, Hugoson A. Periodontal healing after impacted lower third molar surgery: a retrospective study. Int J Oral Surg. 1985;14(1):29-40. 14. Kugelberg CF. Periodontal healing two and four years after impacted lower third molar surgery: a comparative retrospective study. Int J Oral Maxillofac Surg. 1990;19(6):341-345. 15. Kan KW, Liu JK, Lo EC, Corbet EF, Leung WK. Residual periodontal defects distal to the mandibular second molar 6-36 months after impacted third molar extraction. J Clin Periodontol. 2002;29(11): 1004-1011. 16. Osborne WH, Snyder AJ, Tempel TR. Attachment levels and crevicular depths at the distal of mandibular second molars following removal of adjacent third molars. J Periodontol. 1982;53(2):93-95. 17. Ferreira CE, Grossi SG, Novaes AB Jr, Dunford RG, Feres-Filho EJ. Effect of mechanical treatment on healing after third molar extraction. Int J Periodontics Restorative Dent. 1997;17(3):250-259. 18. Leung WK, Corbet EF, Kan KW, Lo EC, Liu JK. A regimen of systematic periodontal care after removal of impacted mandibular third molars manages periodontal pockets associated with the mandibular second molars. J Clin Periodontol. 2005;32(7):725-731. 19. Rosa AL, Carneiro MG, Lavrador MA, Novaes AB Jr. Influence of flap design on periodontal healing of second molars after extraction of impacted mandibular third molars. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2002;93(4):404-407. 20. Kirtiloglu T, Bulut E, Sumer M, Cengiz I. Comparison of 2 flap designs in the periodontal healing of second molars after fully impacted mandibular third molar extractions. J Oral Maxillofac Surg. 2007;65(11): 2206-2210. 21. Pecora GC, Celletti R, Davarpanah M, Covani U, Etienne D. The effects of guided tissue regeneration on healing after impacted mandibular third-molar surgery: 1-year results. Int J Periodontics Restorative Dent. 1993; 13(5):396-407. 22. Aimetti M, Pigella E, Romano F. Clinical and radiographic evaluation of the effects of guided tissue regeneration using resorbable membranes after extraction of impacted mandibular third molars. Int J Periodontics Restorative Dent. 2007;27(1):51-59. 23. Wang HL, Greenwell H, Fiorellini J, et al. Periodontal regeneration. J Periodontol. 2005;76(9):1601-1622. 24. Susin C, Wikesjo UM. Regenerative periodontal therapy: 30 years of lessons learned and unlearned. Periodontol 2000. 2013;62(1): 232-242. 25. Murphy KG, Gunsolley JC. Guided tissue regeneration for the treatment of periodontal intrabony and furcation defects: a systematic review. Ann Periodontol. 2003;8(1):266-302. 26. Sculean A, Nikolidakis D, Nikou G, Ivanovic A, Chapple IL, Stavropoulos A. Biomaterials for promoting periodontal regeneration in human intrabony defects: a systematic review. Periodontol 2000. 2015;68(1): 182-216. 27. Oxford GE, Quintero G, Stuller CB, Gher ME. Treatment of 3rd molar-induced periodontal defects with guided tissue regeneration. J Clin Periodontol. 1997;24(7):464-469. 28. Karapataki S, Hugoson A, Kugelberg CF. Healing following GTR treatment of bone defects distal to mandibular 2nd molars after surgical removal of impacted 3rd molars. J Clin Periodontol. 2000;27(5): 325-332.

10 JADA

( )

- -

http://jada.ada.org

-

2016

29. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343:d5928. 30. Centre for Evidence-based Medicine. Oxford Centre for Evidencebased Medicine: levels of evidence (March 2009). Available at: www.cebm. net/oxford-centre-evidence-based-medicine-levels-evidence-march-2009/. Accessed March 20, 2016. 31. Follmann D, Elliott P, Suh I, Cutler J. Variance imputation for overviews of clinical trials with continuous response. J Clin Epidemiol. 1992; 45(7):769-773. 32. Imputing standard deviations for changes from baseline (Chapter 16. 1.3.2). In: Higgins J, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 (updated March 2011). The Cochrane Collaboration; 2011. Available at: http://handbook.cochrane.org/. Accessed March 20, 2016. 33. Obtaining standard deviations from standard errors, confidence intervals, t values and P values for differences in means (Chapter 7.7.3.3). In: Higgins J, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 (updated March 2011). The Cochrane Collaboration; 2011. Available at: http://handbook.cochrane.org/. Accessed March 20, 2016. 34. Heterogeneity (Chapter 9.5). In: Higgins J, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 (updated March 2011). The Cochrane Collaboration; 2011. Available at: http:// handbook.cochrane.org/. Accessed March 20, 2016. 35. Throndson RR, Sexton SB. Grafting mandibular third molar extraction sites: a comparison of bioactive glass to a nongrafted site. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2002;94(4):413-419. 36. Dodson TB. Management of mandibular third molar extraction sites to prevent periodontal defects. J Oral Maxillofac Surg. 2004;62(10): 1213-1224. 37. Sammartino G, Tia M, Bucci T, Wang HL. Prevention of mandibular third molar extraction-associated periodontal defects: a comparative study. J Periodontol. 2009;80(3):389-396. 38. Hassan KS, Marei HF, Alagl AS. Does grafting of third molar extraction sockets enhance periodontal measures in 30- to 35-year-old patients? J Oral Maxillofac Surg. 2012;70(4):757-764. 39. Tabrizi R, Khorshidi H, Shahidi S, Gholami M, Kalbasi S, Khayati A. Use of lincomycin-impregnated demineralized freeze-dried bone allograft in the periodontal defect after third molar surgery. J Oral Maxillofac Surg. 2014;72(5):850-857. 40. Identifying and measuring heterogeneity (Chapter 9.5.2). In: Higgins J, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 (updated March 2011). The Cochrane Collaboration; 2011. Available at: http://handbook.cochrane.org/. Accessed March 20, 2016. 41. Reynolds MA, Aichelmann-Reidy ME, Branch-Mays GL, Gunsolley JC. The efficacy of bone replacement grafts in the treatment of periodontal osseous defects: a systematic review. Ann Periodontol. 2003;8(1): 227-265. 42. Kao RT, Nares S, Reynolds MA. Periodontal regeneration—intrabony defects: a systematic review from the AAP Regeneration Workshop. J Periodontol. 2015;86(2 suppl):S77-S104. 43. Laurell L, Gottlow J, Zybutz M, Persson R. Treatment of intrabony defects by different surgical procedures: a literature review. J Periodontol. 1998;69(3):303-313. 44. Becker W, Becker BE. Treatment of mandibular 3-wall intrabony defects by flap debridement and expanded polytetrafluoroethylene barrier membranes: long-term evaluation of 32 treated patients. J Periodontol. 1993; 64(11 suppl):1138-1144. 45. Selvig KA, Kersten BG, Wikesjo UM. Surgical treatment of intrabony periodontal defects using expanded polytetrafluoroethylene barrier membranes: influence of defect configuration on healing response. J Periodontol. 1993;64(8):730-733. 46. Steffensen B, Suzuki H, Caffesse RG, Ash MM. Repair of periodontal angular bony defects evaluated by one- and two-dimensional radiographic analysis. Oral Surg Oral Med Oral Pathol. 1987;63(1): 109-114. 47. Cortellini P, Paolo G, Prato P, Tonetti MS. Long-term stability of clinical attachment following guided tissue regeneration and conventional therapy. J Clin Periodontol. 1996;23(2):106-111. 48. Cortellini P, Tonetti MS. Long-term tooth survival following regenerative treatment of intrabony defects. J Periodontol. 2004;75(5):672-678.

ORIGINAL CONTRIBUTIONS

49. Stavropoulos A, Karring T. Guided tissue regeneration combined with a deproteinized bovine bone mineral (Bio-Oss) in the treatment of intrabony periodontal defects: 6-year results from a randomized-controlled clinical trial. J Clin Periodontol. 2010;37(2):200-210. 50. Silvestri M, Rasperini G, Milani S. 120 infrabony defects treated with regenerative therapy: long-term results. J Periodontol. 2011;82(5): 668-675.

51. Marmary Y, Brayer L, Tzukert A, Feller L. Alveolar bone repair following extraction of impacted mandibular third molars. Oral Surg Oral Med Oral Pathol. 1986;61(4):324-326. 52. Kugelberg CF, Ahlstrom U, Ericson S, Hugoson A, Thilander H. The influence of anatomical, pathophysiological and other factors on periodontal healing after impacted lower third molar surgery; a multiple regression analysis. J Clin Periodontol. 1991;18(1):37-43.

JADA

( )

- -

http://jada.ada.org

-

2016 11

ORIGINAL CONTRIBUTIONS

Appendix 1 PUBMED SEARCH STRATEGY

(“Molar, Third”[MH] OR Third molar*[TIAB] OR 3rd molar*[TIAB] OR wisdom tooth [TIAB] OR wisdom teeth [TIAB]) AND (“Tooth Extraction”[MH] OR extract*[TIAB] OR remov*[TIAB]) AND (“Bone Transplantation”[MH] OR “Bone Substitutes”[MH] OR “Transplantation, Heterologous”[MH] OR “Guided Tissue Regeneration, Periodontal”[MH] OR graft* [TIAB] OR xenograft*[TIAB] OR allograft*[TIAB] OR alloplast*[TIAB] OR bone filler*[TIAB] OR bio oss [TIAB] OR “Membranes, Artificial”[MH] OR “Polytetrafluoroethylene”[MH] OR membrane*[TIAB] OR barrier*[TIAB] OR eptfe[TIAB] OR polytetrafluoroethylene[TIAB] OR biomend[TIAB] OR bio gide[TIAB] OR “Bone Morphogenetic Proteins”[MH] OR bone morphogenic*[TIAB] OR bmp*[TIAB] OR “PlateletDerived Growth Factor”[MH] OR platelet derived* [TIAB] OR pdgf[TIAB] OR “Dental Enamel Proteins”[MH] OR enamel[TIAB] OR emdogain[TIAB] OR “Blood Platelets”[MH] OR platelet*[TIAB]) MH stands for MeSH Terms. TIAB stands for Title and Abstract. Words and numbers included in a citation’s title, collection title, abstract, other abstract, and key words. English language abstracts are taken directly from the published article. If an article does not have a published abstract, the U.S. National Library of Medicine (NLM) does not create one. EMBASE SEARCH STRATEGY

(((third OR 3rd) NEXT/1 molar*):ab,ti OR ‘wisdom tooth’:ab,ti OR ‘wisdom teeth’:ab,ti) AND (‘tooth extraction’/exp OR extract*:ab,ti OR remov*:ab,ti) AND (‘bone transplantation’/exp OR ‘bone prosthesis’/exp OR ‘xenograft’/exp OR graft*:ab,ti OR xenograft*:ab,ti OR allograft*:ab,ti OR alloplast*:ab,ti OR (bone NEXT/1 filler*):ab,ti OR ‘bio oss’:ab,ti OR ‘artificial membrane’/ exp OR ‘politef implant’/exp OR ‘politef’/exp OR membrane*:ab,ti OR barrier*:ab,ti OR eptfe:ab,ti OR polytetrafluoroethylene:ab,ti OR politef:ab,ti OR biomend:ab,ti OR ‘bio gide’:ab,ti OR ‘bone morphogenetic protein’/exp OR (bone NEXT/1 morphogenic*):ab,ti OR bmp*:ab,ti OR ‘platelet derived growth factor’/ exp OR (platelet NEXT/1 derived*):ab,ti OR pdgf:ab,ti OR ‘enamel protein’/exp OR enamel:ab,ti OR emdogain:ab,ti OR ‘thrombocyte’/exp OR platelet*:ab,ti OR thrombocyte*:ab,ti) WEB OF SCIENCE SEARCH STRATEGY

Science Citation Index Expanded (SCI-EXPANDED): 1900-present. Conference Proceedings Citation Index-Science (CPCI-S): 1990-present. Book Citation Index–Science (BKCI-S): 2005-present.

11.e1 JADA

( ) http://jada.ada.org

- -

-

2016

TS¼(“Third molar*” OR “3rd molar*” OR “wisdom tooth” OR “wisdom teeth”) AND TS¼(extract* OR remov*) AND TS¼(graft* OR xenograft* OR allograft* OR alloplast* OR “bone filler*” OR “bio oss” OR membrane* OR barrier* OR eptfe OR polytetrafluoroethylene OR politef OR biomend OR “bio gide” OR “bone morphogenic*” OR bmp* OR “platelet derived*” OR pdgf OR enamel OR emdogain OR platelet* OR thrombocyte*) DENTAL AND ORAL SCIENCES SOURCE SEARCH STRATEGY

(“Third molar*” OR “3rd molar*” OR “wisdom tooth” OR “wisdom teeth”) AND (extract* OR remov*) AND (graft* OR xenograft* OR allograft* OR alloplast* OR “bone filler*” OR “bio oss” OR membrane* OR barrier* OR eptfe OR polytetrafluoroethylene OR politef OR biomend OR “bio gide” OR “bone morphogenic*” OR bmp* OR “platelet derived*” OR pdgf OR enamel OR emdogain OR platelet* OR thrombocyte*) Appendix 2 QUALITY ASSESSMENT OF THE RANDOMIZED CONTROLLED TRIALS

Randomization: The method of generating the allocation sequence had to be described in sufficient detail and properly conducted to earn “low risk of bias” in this assessment item. Allocation concealment: Participants and investigators enrolling participants could not foresee assignment before the randomization process to earn “low risk of bias” in this assessment item. Masking of patients: Owing to the difficulty in masking the surgeon performing the procedure without notice, we did not assess the “masking of surgeons.” The studies had to clearly provide the information about the masking of patients before, during, and after the procedures to be assigned to the “low risk of bias” category. Masking of outcome assessment: If the masked or independent examiners had collected and measured clinical outcomes, we assigned the study a “low risk of bias” in this assessment item. Incomplete outcome data and selective reporting: There was no evidence of incomplete outcome data or selective reporting in any study. Therefore, we assigned “low risk of bias” to these studies. Group imbalance: The baseline characteristics of each group in the studies were generally homogeneous. We assigned “low risk of bias” to these studies. Sample size: Sample size had to be decided by the power calculation to earn a “low risk of bias” assignment in this assessment. No included study explained the choice of the number of patients; therefore, we assigned “high risk of bias” to every study in this category.

ORIGINAL CONTRIBUTIONS

Clinician bias: If only 1 surgeon performed the surgical procedures, we assigned the study to have a “low risk of bias” in this assessment item. Appendix 3 EXCLUDED STUDIES

After the authors screened the abstracts and titles of 1,083 articles in the initial search, they determined that 30 studies were eligible for full-text review. The authors excluded 23 studies because the articles did not fit the selection criteria. The reasons for excluding studies were as follows: - Incomplete or no data of clinical outcomes of interest (14 studies) - No control group (4 studies) - The follow-up period was shorter than 6 months (2 studies) - The studies shared the same patient cohort (2 studies) - Insufficient number of patients (1 studies) The authors found 7 prospective studies to be eligible for the systematic review. Twenty-three studies were excluded.e1-e23 Supplementary References e1. Aimetti M, Romano F. Use of resorbable membranes in periodontal defects treatment after extraction of impacted mandibular third molars. Minerva Stomatol. 2007;56(10):497-508. e2. Arenaz-Bua J, Luaces-Rey R, Sironvalle-Soliva S, et al. A comparative study of platelet-rich plasma, hydroxyapatite, demineralized bone matrix and autologous bone to promote bone regeneration after mandibular impacted third molar extraction. Med Oral Patol Oral Cir Bucal. 2010;15(3):e483-e489. e3. Baslarli O, Tumer C, Ugur O, Vatankulu B. Evaluation of osteoblastic activity in extraction sockets treated with platelet-rich fibrin. Med Oral Patol Oral Cir Bucal. 2015;20(1):e111-e116. e4. Celio-Mariano R, de Melo WM, Carneiro-Avelino C. Comparative radiographic evaluation of alveolar bone healing associated with autologous platelet-rich plasma after impacted mandibular third molar surgery. J Oral Maxillofac Surg. 2012;70(1):19-24. e5. Coceancig PL. Alveolar bone grafts distal to the lower second molar. J Maxillofac Oral Surg. 2009;8(1):22-26. e6. Corinaldesi G, Lizio G, Badiali G, Morselli-Labate AM, Marchetti C. Treatment of intrabony defects after impacted mandibular third molar removal with bioabsorbable and non-resorbable membranes. J Periodontol. 2011;82(10):1404-1413. e7. Cortell-Ballester I, Figueiredo R, Valmaseda-Castellon E, Gay-Escoda C. Effects of collagen resorbable membrane placement after the surgical extraction of impacted lower third molars. J Oral Maxillofac Surg. 2015;73(8):1457-1464.

e8. Dodson TB. Reconstruction of alveolar bone defects after extraction of mandibular third molars: a pilot study. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 1996;82(3):241-247. e9. Dodson TB. Is there a role for reconstructive techniques to prevent periodontal defects after third molar surgery? J Oral Maxillofac Surg. 2005; 63(7):891-896. e10. DurmuSlar MC, Alpaslan C, Alpaslan G, Cakir M. Clinical and radiographic evaluation of the efficacy of platelet-rich plasma combined with hydroxyapatite bone graft substitutes in the treatment of intrabony defects in maxillofacial region. Acta Odontol Scand. 2014;72(8): 948-953. e11. Gawai KT, Sobhana CR. Clinical evaluation of use of platelet rich plasma in bone healing. J Maxillofac Oral Surg. 2015;14(1):67-80. e12. Girish Rao S, Bhat P, Nagesh KS, et al. Bone regeneration in extraction sockets with autologous platelet rich fibrin gel. J Maxillofac Oral Surg. 2013;12(1):11-16. e13. Karapataki S, Hugoson A, Falk H, Laurell L, Kugelberg CF. Healing following GTR treatment of bone defects distal to mandibular 2nd molars using resorbable and non-resorbable barriers. J Clin Periodontol. 2000; 27(5):333-340. e14. Kaul RP, Godhi SS, Singh A. Autologous platelet rich plasma after third molar surgery: a comparative study. J Maxillofac Oral Surg. 2012;11(2): 200-205. e15. Kaur P, Maria A. Efficacy of platelet rich plasma and hydroxyapatite crystals in bone regeneration after surgical removal of mandibular third molars. J Maxillofac Oral Surg. 2013;12(1):51-59. e16. Kumar N, Prasad K, Ramanujam L, K R, Dexith J, Chauhan A. Evaluation of treatment outcome after impacted mandibular third molar surgery with the use of autologous platelet-rich fibrin: a randomized controlled clinical study. J Oral Maxillofac Surg. 2015;73(6):1042-1049. e17. Munhoz EA, Ferreira Junior O, Yaedu RY, Granjeiro JM. Radiographic assessment of impacted mandibular third molar socket filled with composite xenogenic bone graft. Dentomaxillofac Radiol. 2006;35(5): 371-375. e18. Ogundipe OK, Ugboko VI, Owotade FJ. Can autologous platelet-rich plasma gel enhance healing after surgical extraction of mandibular third molars. J Oral Maxillofac Surg. 2011;69(9):2305-2310. e19. Oxford GE, Quintero G, Stuller CB, Gher ME. Treatment of 3rd molar-induced periodontal defects with guided tissue regeneration. J Clin Periodontol. 1997;24(7):464-469. e20. Panday V, Upadhyaya V, Berwal V, et al. Comparative evaluation of G bone (hydroxyapatite) and G-graft (hydroxyapatite with collagen) as bone graft material in mandiublar III molar extracction socket. J Clin Diagn Res. 2015;9(3):ZC48-ZC52. e21. Ruga E, Gallesio C, Boffano P. Platelet-rich fibrin and piezoelectric surgery: a safe technique for the prevention of periodontal complications in third molar surgery. J Craniofac Surg. 2011;22(5):1951-1955. e22. Rutkowski JL, Johnson DA, Radio NM, Fennell JW. Platelet rich plasma to facilitate wound healing following tooth extraction. J Oral Implantol. 2010;36(1):11-23. e23. Singh M, Bhate K, Kulkarni D, Santhosh Kumar SN, Kathariya R. The effect of alloplastic bone graft and absorbable gelatin sponge in prevention of periodontal defects on the distal aspect of mandibular second molars, after surgical removal of impacted mandibular third molar: a comparative prospective study. J Maxillofac Oral Surg. 2015;14(1): 101-106.

JADA

( )

- -

http://jada.ada.org

-

2016 11.e2

ORIGINAL CONTRIBUTIONS

eTABLE

Quality assessment of the selected studies. PECORA AND COLLEAGUES,21 1993

THRONDSON AND SEXTON,35 2002

DODSON,36 2004

AIMETTI AND COLLEAGUES,22 2007

SAMMARTINO AND COLLEAGUES,37 2009

HASSAN AND COLLEAGUES,38 2012

TABRIZI AND COLLEAGUES,39 2014

Low

Unclear

Unclear

Unclear

Low

Low

Unclear

Allocation Concealment

Unclear

Unclear

Unclear

Low

Low

Unclear

Unclear

Masking of Patients

Unclear

Unclear

Low

Unclear

Unclear

Low

Low

Masking, Independent of Outcome Assessment

Unclear

Unclear

High

Low

Low

Unclear

Low

Incomplete Outcome Data

Low

Low

Low

Low

Low

Low

Low

Selective Reporting

Low

Low

Low

Low

Low

Low

Low

BIAS

Random Sequence Generation

Other Sources of Bias Group imbalance

Low

Low

Low

Low

Low

Low

Low

Sample size

High

High

High

High

High

High

High

Clinician bias

Unclear

Unclear

Low

Low

Low

Low

Low

No. of Low Risk of Bias

4

3

5

6

7

6

6

Evidence Level*

2b

2b

2b

2b

1b

2b

2b

* The evidence level was based on the Oxford Centre for Evidence-based Medicine30 Levels of Evidence 1.

11.e3 JADA

( ) http://jada.ada.org

- -

-

2016

ORIGINAL CONTRIBUTIONS

STANDARD ERROR OF THETA

0

.5

1

1.5 0

1

2

3

4

5

THETA

eFigure 1. Funnel plot of weighted mean difference of clinical attachment level gain.

STANDARD ERROR OF THETA

0

.5

1

1.5 –2

0

2

4

THETA

eFigure 2. Funnel plot of weighted mean difference of probing depth reduction.

JADA

( )

- -

http://jada.ada.org

-

2016 11.e4