The aim of this meta-analysis was to analyse the effectiveness of submucosal injection of dexamethasone to control the postoperative signs and symptoms resulting from impacted third molar surgery. An electronic search was conducted, without restriction on date or language, in the PubMed/MEDLINE, Cochrane Central Register of Controlled Trials, and Web of Science databases until June 2015 . The eligibility criteria included non-randomized or randomized clinical trials in humans. After the search and selection process, eight articles were included. The fixed-effects or random-effects model, depending on heterogeneity, was built on the inverse-variance method used. The estimations of intervention were expressed as the mean difference (MD) in millimetres. The results of this meta-analysis suggest that the submucosal injection of dexamethasone presents a reduction in the postoperative signs and symptoms resulting from impacted third molar surgery, especially those associated with oedema and pain. In relation to trismus, the meta-analysis showed no statistically significant difference between dexamethasone and the placebo solution.
The surgical extraction of impacted third molars is the most frequent surgical intervention in dentistry. This procedure is often accompanied by swelling, pain, and trismus, which are physiological events involved in the inflammatory response; if exacerbated, these can interfere with the patient’s comfort and social life. After injury to the tissues, a cascade of inflammatory responses is initiated, with the release of chemical mediators such as prostaglandins and leukotrienes and a consequent increase in permeability and local vascularity. The intensity of the inflammatory response will vary according to each patient’s physiology.
As a strategy to minimize the inflammatory response, quick and well-planned procedures performed by experienced professionals are necessary. In addition, postoperative drug therapy control of the inflammation is an effective measure. Of the numerous pharmacological options available, corticosteroids are a potent modulator of inflammation, acting by inhibiting phospholipase A 2 , a chemical mediator that is responsible for the induction of arachidonic acid. With a decrease in this enzyme, there will be less stimulation of immune cells, such as neutrophils, to release prostaglandins and leukotrienes, and consequently there will be a lower intensity of postoperative signs and symptoms.
Various studies have examined the influence of systematically administered corticosteroids before or after the extraction of third molars, with good results observed. Dexamethasone has been employed for years in oral surgery due to its powerful mechanism of action and prolonged half-life. Several protocols for the administration of dexamethasone in third molar surgery have been proposed; however a consensus has yet to be reached. An interesting option that has yielded good results is the administration of dexamethasone through the submucosa. Factors such as the low dosage, localized administration, and short-term exposure generate minimal adverse effects, for example immunosuppression.
The aim of this meta-analysis was to analyse the effectiveness of submucosal injection of dexamethasone to control the postoperative signs and symptoms following impacted third molar surgery.
Materials and methods
The methodology of this review followed the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions and the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses). The clinical questions were broken down and organized using the PICOS strategy (Population, Intervention, Comparison, Outcome, and Study design).
The aim of this meta-analysis was to analyse the effectiveness of submucosal injection of dexamethasone to control the postoperative signs and symptoms (oedema, trismus, and pain) after impacted third molar surgery.
An electronic search was conducted, without restriction on date or language, in the PubMed/MEDLINE, Cochrane Central Register of Controlled Trials, and Web of Science databases until June 2015. The search strategy and the PICOS framework can be viewed in Table 1 . In addition, the references lists of prospective studies for inclusion were accessed to search for further articles.
|Population||#1||(Third molar [MeSH] OR Third molars[MeSH] OR Impacted third molar*[all fields])|
|Intervention||#2||(Third molar surgery*[all fields] OR dexamethasone [MeSH] OR submucosal dexamethasone*[all fields] OR submucosal injection*[all fields] OR dexamethasone postoperative*[all fields])|
|Comparisons||#3||(Dexamethasone [MeSH] OR placebo effect [MeSH])|
|Outcomes||#4||Postoperative oedema*[all fields] OR postoperative pain*[all fields] OR postoperative trismus*[all fields]|
|Study design||Randomized controlled trials and controlled clinical trials|
|Search combination||#1 AND #2 AND #3 AND #4|
|Electronic databases||PubMed/MEDLINE, Cochrane Central Register of Controlled Trials, and Web of Science|
This review sought randomized controlled trials (RCTs) and prospective controlled studies comparing the effect of the submucosal injection of dexamethasone with that of placebo after impacted third molar surgery in humans. The exclusion criteria were animal studies, retrospective cohort studies, clinical series, case reports, and reviews. Studies involving volunteers with decompensated metabolic diseases were also excluded.
The search and screening process was conducted by two authors/reviewers (VMF and EPB), who first analyzed the titles and abstracts. At a second stage, full articles were selected for careful reading and were analyzed according to the eligibility criteria (inclusion/exclusion) for future data extraction. Disagreements between reviewers were settled through careful discussion. The concordance between the two reviewers in relation to the search procedure was evaluated by Cohen’s kappa (κ) statistical test. If necessary, the authors of the studies were contacted by e-mail to answer any questions.
The quality analysis of the RCTs was conducted using the Cochrane Collaboration tool for assessing risk of bias in RCTs . The analysis of each study was based on six criteria: sequence generation (Was the allocation sequence adequately generated?), allocation concealment (Was allocation adequately concealed?), blinding (Was knowledge of the allocated intervention adequately prevented during the study?), incomplete outcome data (Were incomplete outcome data adequately addressed?), selective outcome reporting (Are reports of the study free of suggestions of selective outcome reporting?), and other sources of bias (Was the study apparently free of other problems that could put it at a high risk of bias?).
The following data were extracted from the studies included (when available): authors, year of publication, study design, number of subjects, mean age, sex, number of smokers, jaw, impaction type (Pell and Gregory ), mean duration of surgery, dexamethasone dosage, use of antibiotics, use of mouth rinse, time of measurement, method of assessment of oedema, pain, and trismus, and P -values.
Continuous variables (oedema, pain, and trismus) from the studies included were subjected to meta-analysis when at least two studies analysed the same data type. The estimation of the intervention effect was expressed as the mean difference (MD) in millimetres with a 95% confidence interval (CI). The inverse-variance method was used for the random-effects model or the fixed-effects model. The I 2 statistic was used to express the percentage of heterogeneity of the studies. Values of up to 25% were classified as low heterogeneity, values of 50% as medium heterogeneity, and values of >70% as high heterogeneity. When a significant heterogeneity was found ( P < 0.10), the results of the random-effects model were validated. When low heterogeneity was verified, the fixed-effects model was considered. The level of statistical significance was set at P < 0.05.
Publication bias was explored graphically by means of a funnel plot. An asymmetric funnel plot may indicate possible publication bias.
The data were analyzed using the statistical software Review Manager version 5.2.8 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark; 2014).
The initial search resulted in 656 titles from PubMed/MEDLINE, 83 titles from Cochrane Central Register of Controlled Trials, and 120 from Web of Science. After the first assessment, 13 full-text articles were selected. After careful reading, five studies were excluded as they failed to conform to the eligibility criteria of this review. Therefore, eight studies, published between 2006 and 2014, were included in this systematic review. The reasons for exclusion of the studies and the selection process are depicted in Fig. 1 .
The κ concordance value for the two reviewers was 0.79 for the potential articles to be included (titles and abstracts) and 0.85 for the articles selected; this indicates substantial concordance for the potential articles and ‘almost perfect’ concordance for the articles selected, according to the criteria proposed by Landis and Koch.
The characteristics of the studies included are presented in Table 2 . The number of participants in the studies ranged from 30 to 100. A total 476 patients were analysed, 256 in the test group and 220 in the control group. All of the surgical procedures were performed on the lower third molars, with an average surgery time of 35.3 min for the test group and 38 min for the control group. Four articles reported the number of smokers included. One study failed to mention the selection criteria (classification) for the impaction type of the third molars included in the study. All other works adopted the Pell and Gregory classification.
|Authors (year)||Study design
No. of subjects
No. per group
No. of smokers
(Pell and Gregory)
|Mean duration of surgery, min||Dexamethasone dosage, mg||Antibiotics/mouth rinse, days|
|Graziani et al. (2006)||RCT
14 G1, 14 G2, 15 G3 a
13 M/30 F
|Mandible||Class 1, 2, or 3 and position A or B||G1: 23.2 ± 15.2
G2: 23.9 ± 13.2
G3: 22.3 ± 9.9
|4 and 10||5 or 3/NR|
|Grossi et al. (2007)||RCT
18 G1, 20 G2, 23 G3 a
33 M/28 F
|Mandible||Class 2 and position B||G1: 27.4 ± 9.5
G2: 26.4 ± 9.5
G3: 27.7 ± 11.0
|4 and 8||Prophylactic 1 h before surgery/7|
|Deo and Shetty (2011)||RCT
19 G1, 11 G2
18 M/12 F
|Mandible||Class 1 or 2 and position A, B, or C||G1: 41.3
11 G1, 11 G2 b
9 M/13 F
|Mandible||Class 2 or 3 and position B or C||G1: 36.5 ± 8.5
G2: 31.0 ± 9.6
|Majid and Mahmood (2011)||RCT
10 G1, 10 G2 b
9 M/11 F
|Mandible||Class 2 or 3 and position A, B, or C||G1: 29.6 ± 8.3
G2: 37 ± 8.58
|Nair et al. (2013)||RCT
50 G1, 50 G2
|Mandible||NR||NR||4||Prophylactic 1 h before surgery/NR|
|Warraich et al. (2013)||RCT
50 G1, 50 G2
72 M/28 F
|Mandible||Class NR and position B or C||G1: 69.2 ± 18.8
G2: 66.3 ± 17
|Ehsan et al. (2014)||RCT
50 G1, 50 G2
|Mandible||Class 2 and position B||G1: 40.6 ± 4.8
G2: 42.1 ± 4.54
The studies adopted dexamethasone concentrations of 4 mg, 8 mg, or 10 mg for an oral or lingual submucosal injection. For the control group, all studies used a saline solution as the placebo medication. The submucosal injection of dexamethasone or placebo was performed before the surgical procedure in four studies. Antibiotic medication was administered prophylactically before surgery, or by continuous use after the procedures. Three studies advised participants to perform mouth rinsing with chlorhexidine mouthwash postoperatively.
The results of the quality assessment of the RCTs included are presented in Table 3 . Only one study met all the criteria described in the Cochrane Collaboration tool for assessing risk of bias in RCTs. No articles cited had followed the recommendations of the CONSORT statement for conducting the RCT.
|Authors (year)||Adequate sequence generation||Allocation concealment||Blinding||Incomplete outcome data addressed||Selective outcome reporting||Free of other sources of bias|
|Graziani et al. (2006)||Yes||Yes||Yes||Yes||Yes||Yes|
|Grossi et al. (2007)||Unclear||Yes||Unclear||Yes||Yes||Unclear|
|Deo and Shetty (2011)||Unclear||Unclear||Unclear||Yes||Yes||Unclear|
|Majid and Mahmood (2011)||Yes||Yes||Partial||Yes||Yes||Unclear|
|Nair et al. (2013)||Yes||Yes||Unclear||Yes||Yes||Unclear|
|Warraich et al. (2013)||Unclear||Unclear||Unclear||Yes||Yes||Unclear|
|Ehsan et al. (2014)||Unclear||Unclear||Unclear||Yes||Yes||Unclear|
For the assessment of oedema, the studies performed measurements of the facial contour using pre-established points of reference ( Table 4 ). Six studies provided data appropriate for meta-analysis. The random-effects model was used given the considerable heterogeneity found ( I 2 = 87%; P < 0.00001). The meta-analysis presented a MD of −2.20 (95% CI −2.70 to −1.70), with a statistically significant difference in favour of dexamethasone ( P < 0.00001) ( Fig. 2 ).
|Authors (year)||Day of measurement (early)||Method of assessment||Comparison groups, mm||P -value|
|Graziani et al. (2006)||2||Facial measurements: from the corner of the mouth to the attachment of the ear lobe following the bulge of the cheek and from the outer canthus of the eye to the angle of the mandible||NR
|Grossi et al. (2007)||2||Facial measurements: from tragus to the corner of the mouth and from tragus to pogonion||G1: 2.7 ± 0.1
G2: 5.0 ± 0.1
G3: 7.9 ± 0.45
|Deo and Shetty (2011)||2||Facial measurements: from tragus to corner of mouth, tragus to pogonion, and mandibular angle to ala of the nose||G1: 3.96 ± 1.33
G2: 8.47 ± 2.27
|Majid (2011)||3||Facial measurements: from tragus to midline and from gonion to lateral canthus using a tape measure||G1: 0.5 ± 0.5
G2: 2.3 ± 0.9
|Majid and Mahmood (2011)||3||Facial measurements: from tragus to midline and from gonion to lateral canthus using a tape measure||G1: 0.5 ± 0.6
G2: 2.4 ± 0.9
|Nair et al. (2013)||2||Facial measurements: tragus, pogonion, and the corner of the mouth||G1: 111 ± 5.79
G2: 115 ± 6.61
|Warraich et al. (2013)||2||Facial measurements: from tragus to midline and from gonion to lateral canthus using a 3D optical scanner||NR
|Ehsan et al. (2014)||2||Facial measurements: from earlobe to corner of the mouth and from outer canthus to angle of the mandible||G1: 11.3 ± 0.6
G2: 12.9 ± 0.79
To assess trismus, all studies measured the inter-incisal distance at maximum opening before and after surgery ( Table 5 ). Six studies provided data appropriate for meta-analysis. The random-effects model was used given the considerable heterogeneity found ( I 2 = 94%; P < 0.00001). The meta-analysis presented a MD of −2.92 (95% CI −7.13 to 1.29) and showed no statistically significant difference between the groups analysed ( P = 0.17) ( Fig. 3 ).