Abstract
The extraction of an impacted third molar violates the surrounding soft and bony tissues. The surgeon’s access to the tooth, for which there are various surgical approaches, has an important impact on the periodontium of the adjacent second molar. The aim of this review was to analyze the relationships between the different flap techniques and postoperative periodontal outcomes for the mandibular second molars (LM2) adjacent to the impacted mandibular third molars (LM3). An electronic search of MEDLINE and other databases was conducted to identify randomized controlled trials fulfilling the eligibility criteria. To assess the impact of flap design on the periodontal condition, the weighted mean difference of the probing depth reduction (WDPDR) and the weighted mean difference of the clinical attachment level gain (WDCAG) at the distal surface of LM2 were used as the primary outcomes. The results showed that, overall, the different flap techniques had no significant impact on the probing depth reduction (WDPDR −0.14 mm, 95% confidence interval −0.44 to 0.17), or on the clinical attachment level gain (WDCAG 0.05 mm, 95% confidence interval −0.84 to 0.94). However, a subgroup analysis revealed that the Szmyd and paramarginal flap designs may be the most effective in reducing the probing depth in impacted LM3 extraction, and the envelope flap may be the least effective.
Dentists and oral surgeons have long held the consensus that symptomatic or pathological impacted third molars should be extracted. However, in the case of asymptomatic impacted third molars there is still some debate about the need for removal. The American Association of Oral and Maxillofacial Surgeons (AAOMS) suggests a timely, prophylactic extraction to prevent possible future pathological changes around the impactions or adjacent second molars. This argument was recently supported by Nunn et al., who reported a significantly increased risk of a second molar pathology of 4.88-fold with soft tissue impaction of the adjacent third molar and 2.16-fold with bony impaction of the adjacent third molar. Their cohort study examined data collected over 25 years from 416 adult men and also noted that the incidence of a second molar pathology was lowest in subjects lacking an adjacent impacted third molar. Since the impacted third molars might ultimately be extracted, any associated adverse effects would become a disconcerting, troublesome issue for both dentists and patients.
Impacted teeth are not fully erupted into the oral cavity. Therefore, removal procedures can require a surgical flap, ostectomy, and division of the tooth, as well as suture of the soft tissue. The manipulation of the hard and soft tissues surrounding the impacted third molars can cause postoperative pain and swelling, and possibly periodontal disease associated with the adjacent second molars. Postoperative pain and swelling are transient outcomes, but periodontal disease can have a chronic course that severely compromises the stability of the teeth. Periodontal defect formation often causes plaque accumulation and further local inflammatory disease. Therefore, consideration should be given to minimizing tissue damage around the surgical area. This can be accomplished through the intervention technique (flap design, suture, or additional periodontal regenerative therapy) and the perioperative management chosen by the surgeon during the operation. Over the past few decades, investigations into periodontal outcomes in regards to third molar extraction, as well as perioperative hard and soft tissue management techniques, have been reported extensively in the published literature in efforts to improve periodontal outcomes.
The amount of bone loss and destruction of the periodontium perioperatively are primarily related to the surgeon’s attempt to access, elevate, and luxate the tooth successfully. Since a mucoperiosteal flap is designed for adequate visual accessibility and manipulation of the surgical area by release of the soft tissue, it is intuitively thought to be related to soft tissue sequelae. Many studies have examined the effects of various mucoperiosteal flap designs, including the Szmyd flap, envelope flap, triangular (or three-cornered) flap, and modified versions of these flaps.
For impacted mandibular third molars (LM3), envelope and triangular flaps are possibly two of the most commonly used flaps by oral surgeons. While both envelope and Szmyd flaps consist of a single horizontal incision and flap elevation by undermining the periosteum, triangular flaps utilize an additional vertical buccal releasing incision. The main advantages of the envelope and Szmyd flaps are the minimal disruption of the vascular supply to the elevated tissue and the ease of wound closure. On the other hand, the triangular flap allows extended reflection of the flap for better visibility and accessibility during ostectomy. Also, the relaxing vertical incision decreases flap tension, and thus the triangular flap is also believed to promote rapid wound healing. Nevertheless, each study has proposed and investigated a flap design that the authors believe to provide better outcomes than other flap designs. With studies proposing various flap designs while detailing the failures of other designs, there has been no clear conclusion on the ideal flap design for LM3 extraction.
Although a descriptive review of this subject has been published, a comprehensive, systematic review and meta-analysis looking at the effects of the different flap designs on postoperative periodontal conditions is lacking. The aim of this systematic review was to analyze the effects of various flap designs used for impacted LM3 extraction on the periodontal outcomes of the adjacent mandibular second molars (LM2). The comparison focused primarily on single horizontal flaps (envelope and Szmyd flaps) and flaps with a vertical releasing incision (triangular flaps).
Materials and methods
This systematic review and subsequent meta-analysis were conducted in accordance with the principles of the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses).
PICO (population, intervention, comparator, outcomes) question
“Do different surgical flap techniques for impacted mandibular third molar extraction affect periodontal healing of the adjacent mandibular second molar during at least 3 months of follow-up?”
Study selection criteria
Only prospective randomized controlled trials (RCTs) that had recruited patients with a clinical diagnosis of impacted LM3, who underwent surgical extraction, were included. The studies had to primarily compare the clinical outcomes of impacted LM3 extractions using different surgical flap techniques (for example, envelope flaps, Szmyd flaps, and triangular flaps). Furthermore, the study had to report the clinical outcomes associated with periodontal healing, including changes in periodontal probing depth (PD) and clinical attachment level (CAL) at the distal surface of the adjacent second molars from baseline to the last measurement. Only original research studies that provided sufficient detail regarding the methods and results to enable the use and adjustment of the data and results were considered. A minimum mean follow-up time of 3 months (12 weeks) from baseline (tooth extraction) to the last clinical assessment was required. The studies were published in English.
Literature search
Electronic searches were conducted through the MEDLINE (PubMed), Embase, Cochrane CENTRAL, and Scopus databases to screen all relevant articles published from inception to April 2016. Appendix 1 lists the search strategies used in these databases.
A further search was performed in the following journals: Journal of Oral and Maxillofacial Surgery , International Journal of Oral and Maxillofacial Surgery , British Journal of Oral and Maxillofacial Surgery , Journal of Cranio-Maxillofacial Surgery , International Journal of Oral Surgery , and Journal of Oral Surgery . Moreover, the reference lists of selected articles were screened to identify additional articles that might fit the selection criteria.
Quality assessment of the included studies
The methodological quality assessment was undertaken using the Cochrane Collaboration risk-of-bias assessment tool for RCTs. This tool addresses the following six domains: randomization, allocation concealment, blinding of patients and assessors, completeness of outcome data, and risk of selective outcome reporting. For each study, an assessment of low or high risk of bias was given. However, if insufficient information was given in the article for such a judgement to be made, the domain was assessed as having an ‘unclear’ risk of bias. The summarized overall risk of bias for each domain across all studies included was tabulated. Two authors (C.T.L. and Y.W.C.) assessed the articles independently, and the inter-investigator agreement was evaluated by calculating the κ correlation coefficient.
Data extraction and synthesis
Data were extracted and recorded independently by two authors using a data abstraction form (L.H. and Y.W.C.). Any instances of inconsistency were discussed. For each study included, the following data were retrieved: (1) year of publication; (2) trial design (split-mouth or parallel groups); (3) patient characteristics (number of participants in each group, age and sex of the participants); (4) details of the impacted tooth characteristics; (5) intraoperative and postoperative procedures (types of surgical flaps, ostectomy, debridement of the extraction socket, and postoperative instructions); (6) follow-up information (time intervals, numbers lost to follow-up); and (7) outcomes of interests (PD and CAL).
The primary outcome was the change in PD from baseline to the last clinical measurement at the distal surface of LM2 adjacent to the extraction site. The change in CAL at the distal surface of LM2 was also evaluated. If available, the values at the distal–buccal site were used in the analyses. Otherwise, the values from other sites of the LM2 distal surface were used. Any instances of inconsistency in data extraction were examined by another author (C.T.L.) to ensure the accuracy.
The reduction in PD was calculated by subtracting the PD at the last measurement from the PD at baseline. The gain in CAL was calculated by subtracting the CAL at the last measurement from the CAL at baseline. The standard deviation of the PD reduction or CAL gain was not reported in some of the studies. The missing standard deviations were imputed by using an assumptive correlation coefficient that equals 0.5.
The weighted mean difference of the PD reduction (WDPDR) between the two groups was calculated by subtracting the PD reduction in the control group (triangular flap group) from the PD reduction in the other group (using a different surgical flap technique, such as Szmyd flap, envelope flap, or other modified flap). The weighted mean difference of the CAL gain (WDCAG) between the two groups was calculated by subtracting the CAL gain in the control group (triangular flap group) from the CAL gain in the other group (using a different surgical flap technique).
Heterogeneity across the studies was assessed by χ 2 test and I 2 statistic. Given the significant heterogeneity between studies ( P < 0.01 by χ 2 test), the random-effects model was adopted. Subgroup analyses were undertaken to investigate the effects of clinical factors thought to influence the outcomes. Estimated effect sizes and 95% confidence intervals (CIs) based on the individual study were displayed using forest plots. Meta-regression was performed to evaluate the correlation between the primary outcome and the different variables. P < 0.05 was considered statistically significant. Potential publication bias for the included studies was examined using the Egger regression asymmetry test and illustrated in a funnel plot. All statistical analyses were performed using Stata 11.2 software (StataCorp LP, College Station, TX, USA).
Results
Study selection and characteristics of the included studies
The search strategy yielded a total of 1340 non-duplicate reports from all databases. Based on the predetermined eligibility criteria, 21 articles were selected for full-text assessment after reviewing the titles and abstracts. Thirteen articles were excluded at the final stage of article selection. The reasons for exclusion were as follows: not related to flap designs or not comparing flaps of interest, not RCT, outcomes of interest not reported or insufficient statistical data.
A flowchart of the search and selection process is displayed in Fig. 1 . A total of eight studies were included for qualitative and quantitative analysis. Six of these had a split-mouth design, where participants had one impacted LM3 extracted using a triangular flap and the other LM3 extracted using a different flap. The remaining two studies were of parallel group design, where participants were allocated randomly to extraction with either one of the flaps.
A total of 224 patients were recruited in these studies, with 35 participants dropping out because of loss to follow-up or postoperative complications. The participants were all in good health and had no known infectious or acute inflammatory contraindications for local surgery. The majority of the studies reported the use of ostectomy during the surgeries; however, one study did not report whether mechanical bone removal was performed. Other detailed applicable data gathered from the selected studies are shown in Table 1 .
| First author (year) | Study design | No. of patients recruited | Age range (mean), years | Characteristics of LM3 impaction | Flap design in the other-flap group (number of teeth) | Flap design in the control group (number of teeth) | Ostectomy | Postoperative management | Dropouts |
|---|---|---|---|---|---|---|---|---|---|
| Arta (2011) | RCT Split-mouth | 20 | 18–26 (NA) | NR | Szmyd (20) | Triangular (20) | NR | Chlorhexidine mouth rinse | 0 |
| Baqain (2012) | RCT Split-mouth | 20 | 18–26 (21.4) | NR | Envelope (19) | Triangular (19) | Yes | Socket debridement and chlorhexidine mouth rinse | 1 |
| Briguglio (2011) | RCT Parallel | 45 | 18–41 (NA) | Mesioangular tilting ≥25° in relation to LM2, <50% crown above the cervical margin of LM2 | Envelope (modified) (30) | Triangular (15) | Yes | Chlorhexidine mouth rinse | 0 |
| Chaves (2008) | RCT Parallel | 20 | 18–27 (21.75) | Semi-impacted with highest portion of the crown being at the level of the occlusal plane of LM2 | Szmyd (modified) (10) | Triangular (10) | Yes | Socket debridement | 0 |
| Kirtiloglu (2007) | RCT Split-mouth | 18 | 16–32 (20.8) | Vertical or mesioangular tilting at 25–75° in relation to LM2 | Szmyd (modified) (18) | Triangular (18) | Yes | Socket debridement and chlorhexidine mouth rinse | 0 |
| Korkmaz (2015) | RCT Split-mouth | 28 | 18–28 (22.43) | Vertical or mesioangular tilting | Envelope (28) | Triangular (28) | Yes | Chlorhexidine mouth rinse | 0 |
| Monaco (2009) | RCT Split-mouth | 12 | 15–19 (16) | LM3 root not completely developed | Envelope (12) | Triangular (12) | Yes | Socket debridement and chlorhexidine mouth rinse | 0 |
| Suarez-Cunqueiro (2003) | RCT Split-mouth | 61 | 17–31 (22) | Completely tissue-covered | Triangular (modified) (27) | Triangular (27) | Yes | Chlorhexidine mouth rinse | 34 |
Risk of bias assessment
The results of the quality assessment are shown in Table 2 ( Appendix 2 describes the quality assessment of the RCTs). These studies had total low risk of bias scores ranging from 1 to 6. The κ correlation coefficient was 0.88 between the two authors (C.T.L. and Y.W.C.).
| First author (year) | Randomization (Selection) | Allocation concealment (Selection) | Blinding of patients (Performance) | Blinding of assessor (Detection) | Incomplete outcome data (Attrition) | Selective reporting (Reporting) | Number of low risk of bias |
|---|---|---|---|---|---|---|---|
| Arta (2011) | Low | Unclear | Unclear | Unclear | Low | Low | 3 |
| Baqain (2012) | Low | Unclear | Unclear | Unclear | Low | Low | 3 |
| Briguglio (2011) | Low | Low | Low | Low | Low | Low | 6 |
| Chaves (2008) | Low | Unclear | Low | Low | Low | Low | 5 |
| Kirtiloglu (2007) | Unclear | Unclear | Unclear | Unclear | Low | Low | 2 |
| Korkmaz (2015) | Low | Low | Unclear | Low | Low | Low | 5 |
| Monaco (2009) | High | High | Unclear | Unclear | Low | Low | 2 |
| Suarez-Cunqueiro (2003) | Low | Unclear | Unclear | Unclear | High | High | 1 |
Characteristics of the surgical flap designs
In each of the trials, the descriptions of the design of the incision and flap were slightly different, but all of the studies used standard triangular flaps in one of their groups: four studies compared a standard triangular flap with an envelope or modified envelope flap, while the other four studies compared a standard triangular flap with a Szmyd, modified Szmyd, or modified triangular flap. The types of surgical flaps used in the studies are illustrated in Fig. 2 .
In the standard triangular flap, the incision is initiated distally, from the mandibular ramus to the distobuccal line angle of LM2, and is continued with a sulcular incision around LM2 to the mesiobuccal line angle. Then a vertical releasing incision is made from the mesiobuccal line angle, curving forward to the mandibular vestibule with a slightly obtuse angle of the flap. It is also possible that a buccal releasing incision is made starting from the distobuccal line angle of LM2 or the lower first molar (LM1) ( Fig. 2 a). A modification of the triangular flap was used in the study by Suarez-Cunqueiro et al., which left a 2-mm gingival collar around the buccal side of LM2 ( Fig. 2 d).
The Szmyd flap utilizes a buccal horizontal extension of a distal linear incision from the distobuccal line angle of the second molar, maintaining a stripe of gingival margin on the buccal surface of the molars ( Fig. 2 b). The modified Szmyd flap described in the study of Kirtiloglu et al. was, however, similar to another modified triangular flap; it kept the attached gingiva in the distobuccal region of LM2, with an extension of the incision made straight down to the buccal vestibule ( Fig. 2 e).
The envelope flap involves a distal linear incision along the external oblique ridge and continues with a sulcular incision around LM2 and LM1, sometimes even mesial to the second premolar. The extension of the sulcular excision varies, and it can be extended to the second premolar if necessary ( Fig. 2 c). The modification of the envelope flap made in the study by Briguglio et al., involved extension of the sulcular incision with a direct horizontal cut through the papillae between the first and second molars ( Fig. 2 f).
Reduction of probing depth and gain in clinical attachment level
PD and CAL data were retrieved from the selected studies for quantitative meta-analysis. In order to compare the impacts of the different flap designs on the postoperative periodontal condition, Szmyd flaps, envelope flaps, and all the other modification flaps were grouped together (named the other-flap group) and compared with standard triangular flaps (control group).
The different surgical flap techniques had no significant impact on the PD reduction after a postoperative period of at least 3 months (WDPDR −0.14 mm, 95% CI −0.44 to 0.17; heterogeneity, I 2 = 83.4%) ( Fig. 3 a) . On removing each study individually from the analysis, no study appeared to significantly affect the outcome of WDPDR. The WDPDR was highest (−0.03 mm, 95% CI −0.31 to 0.24) when the study of Korkmaz et al. was excluded and lowest (−0.20 mm, 95% CI −0.53 to 0.12) when the study of Kirtiloglu et al. was excluded.
With regard to WDCAG, only three studies provided data for comparison. The results also showed no difference in WDCAG between the two groups (WDCAG 0.05 mm, 95% CI −0.84 to 0.94; heterogeneity, 77.4%) ( Fig. 3 b).
Subgroup analysis
The data were categorized into three groups according to the type of flap used in the other-flap group: envelope/modified envelope flaps, Szmyd flaps, and the paramarginal design flaps in which an attached gingival collar around LM2 is maintained. The paramarginal design flap has the possible advantage of gingival margin height maintenance by eliminating the need to detach the facial keratinized gingiva. The modified Szmyd flap used by Kirtiloglu et al. and the modified triangular flap used by Suarez-Cunqueiro et al. in the other-flap group had a paramarginal design.
The results demonstrated different effects on PD reduction for the three subgroups. Compared to the standard triangular flap, both the Szmyd flap and paramarginal design flaps had a larger PD reduction after surgery (Szmyd flaps: WDPDR 0.15 mm, 95% CI −0.13 to 0.44, P = 0.33; heterogeneity, I 2 = 0.0%; paramarginal flaps: WDPDR 0.27 mm, 95% CI −0.07 to 0.61, P = 0.23; heterogeneity, I 2 = 0.0%). In the comparison of the envelope/modified envelope flap versus the standard triangular flap, the PD reduction was larger for the standard triangular flap (WDPDR −0.42 mm, 95% CI −0.87 to 0.03, P = 0.06; heterogeneity, I 2 = 90.3%) ( Fig. 3 c).
Meta-regression
The effect of several clinical variables on WDPDR was evaluated (for example, the length of the follow-up period, baseline mean PD, subgroup flap designs, study designs, and quality bias). The results showed that none of these variables had a significant effect on PD reduction ( Appendix 3 ).
Publication bias
There was no significant publication bias detected for WDPDR (Egger’s test, P = 0.80). The funnel plot demonstrated a symmetric distribution of the data in each study ( Fig. 4 ).
Results
Study selection and characteristics of the included studies
The search strategy yielded a total of 1340 non-duplicate reports from all databases. Based on the predetermined eligibility criteria, 21 articles were selected for full-text assessment after reviewing the titles and abstracts. Thirteen articles were excluded at the final stage of article selection. The reasons for exclusion were as follows: not related to flap designs or not comparing flaps of interest, not RCT, outcomes of interest not reported or insufficient statistical data.
A flowchart of the search and selection process is displayed in Fig. 1 . A total of eight studies were included for qualitative and quantitative analysis. Six of these had a split-mouth design, where participants had one impacted LM3 extracted using a triangular flap and the other LM3 extracted using a different flap. The remaining two studies were of parallel group design, where participants were allocated randomly to extraction with either one of the flaps.
