A meta-analysis of pre- and postoperative corticosteroids for reducing the complications following facial reconstructive and aesthetic surgery

Introduction Edema and ecchymosis after facial plastic surgery are a troublesome concern for both patients and surgeons. Corticosteroid administration is thought to shorten the recovery period and reduce these sequelae. Data regarding the efficacy of corticosteroid administration remains controversial among surgeons. Objective We conducted this systematic review and meta-analysis to determine the effect of pre- and postoperative corticosteroids on postoperative complications in patients undergoing facial reconstructive surgery supported with different subgroup analysis. Methods A comprehensive literature search of articles was conducted in PubMed, Cochrane Central, SCOPUS, and EBSCO through October 2019. We included all clinical trials in which patients underwent any type of facial plastic surgery to study the effect of corticosteroids on postoperative complications. We performed subgroup analysis according to the types and doses of corticosteroid preparation, in addition to a subgroup analysis of pre- or postoperative corticosteroid usage. All statistical analysis was performed using the RevMan software. Results Nineteen studies were included in this systematic review, but only 10 of them were eligible for meta-analysis. The periorbital edema and ecchymosis scores were significantly reduced in the corticosteroids group compared to placebo −0.82, 95% CI (−1.37, −0.26), and -0.95, 95% CI (−1.32, −0.57), respectively. However, these significant differences were not maintained at day 3 and 7. Smaller doses of corticosteroid (8 mg and 10 mg) were associated with smaller differences in the mean score of upper and lower eyelid edema and ecchymosis, while the higher doses were associated with greater differences. Furthermore, preoperative corticosteroid usage significantly reduced the intraoperative bleeding when compared to placebo for higher doses > 50 mg per day (p < 0.0001), but not for 8 mg corticosteroid (p = 0.06). Adding postoperative steroid dose to the preoperative one was associated with less edema and ecchymosis than preoperative administration alone. Conclusion This comprehensive meta-analysis confirms a statistically significant benefit of preoperative corticosteroids. Furthermore, continuing the steroids postoperatively is associated with long-term reduction of complications. Higher doses of corticosteroids are associated with a more significant reduction in edema and ecchymosis, but further studies are recommended to determine the postoperative side effects, including surgical site infection and delayed healing.


Introduction
Facial plastic surgeries are one of the most commonly performed operative procedures worldwide. Facial surgery includes many types of procedures, including facelifts, rhinoplasty, and maxillofacial plastic surgery. 1---3 Edema, ecchymosis and intraoperative bleeding are the most common postoperative complications after these surgeries. 4 The severity of edema and ecchymosis is the most important concern as it can delay the healing process of the involved tissues and alter the final intended aesthetic outcome 5 Various methods and concepts with variable success outcomes were developed to avoid these conditions. 6 Corticosteroid administration is thought to shorten the recovery period and reduce postoperative edema and ecchymosis. 7 It provides anti-inflammatory properties by inhibiting the initial process of inflammation that involves the migration of lymphocytes, fibrin deposition, capillary dilatation, and phagocytic activity. 8 Some studies report that the preoperative administration of steroids is effective, whereas postoperative administration is not. 6,9 Data regarding the efficacy of corticosteroid administration remains controversial among surgeons, with a debate on the long-term benefit of intraoperative or postoperative corticosteroids administration in patients undergoing facial plastic surgery. 10---12 In most of these studies, various doses and routes of administration were used. Griffies et al. were the first to investigate the effect of corticosteroid on the edema and ecchymosis of facial plastic surgery through a prospective randomized clinical trial. 13 They reported that a one-time bolus of dexamethasone 10 mg could decrease the facial swelling significantly when compared to the placebo. In 1991, Hoffmann et al. examined the preoperative and postoperative use of dexamethasone 10 mg, and reported it yielded a significant reduction in postoperative eyelid edema, para/intra-nasal edema and ecchymosis. 14 On the other hand, Berinstein et al. 15 concluded that the rhinoplasty patients who received dexamethasone experienced increased postoperative edema when compared to controls as assessed by magnetic resonance imaging scans.  reported that pre-and postoperative 10 mg dexamethasone with controlled hypotension considerably reduced postoperative swelling and ecchymosis following rhinoplasty, as well as intraoperative bleeding. 16 The Turkish study by Gurlek et al. demonstrated that steroids (betamethasone 8 mg, dexamethasone 8 mg, and methylprednisolone 40 mg were not effective in preventing or reducing edema and ecchymosis after open rhinoplasty with osteotomies, with no differences in the levels of ecchymosis or edema among the steroid groups, the tenoxicam group, and the placebo groups being observed. 2 Recently, in 2018, Sanober et al. showed that preoperative dexamethasone 8 mg could decrease periorbital edema by 50%, compared to placebo 33.3%. By the 7th day postoperatively, only 3.3% in the dexamethasone group had Grade III edema compared to 13.3%in the placebo group.
Considering how common and important facial surgery procedures are, we conducted this systematic review and meta-analysis to resolve the conflicting conclusions reported earlier and to determine the effect of pre-and postoperative corticosteroids on postoperative complications in patients undergoing facial reconstructive surgery.

Methods
We performed all steps of this systematic review in strict accordance with the Cochrane handbook of systematic reviews and meta-analysis. 17 We also followed the Preferred reporting items for systematic reviews and metaanalyses (PRISMA statement guidelines) during drafting our manuscript. 18 Eligibility screening was conducted in two steps, each by two independent reviewers (Saud Abdalwahab Aldhabaan and Jibril Yahya Hudise): a) Title and abstract screening for matching the inclusion criteria, and b) Full-text screening for eligibility for meta-analysis. Disagreements were resolved employing the opinion of a third reviewer (Amani Obeid).

Data extraction
Two independent reviewers (Saud Abdalwahab Aldhabaan and Amani Obeid) extracted the data and another reviewer (Jibril Yahya Hudise) resolved disagreements. The extracted data included the following: a) General characteristics of each study including; study setting, study design, sample size, type of intervention and the doses, type of competitor; b) Patients' baseline characteristics of each study including: age, gender, race; c) Operation characteristics including: type of surgery, duration of surgery, follow-up period, and postoperative edema assessment; d) Outcomes of interest including swelling (edema), bruising (ecchymosis), intraoperative bleeding and e) Risk of bias criteria.

Risk of bias assessment
To assess the risk of bias within the included clinical trial, two independent reviewers (Saud Abdalwahab Aldhabaan and Jibril Yahya Hudise) used the Cochrane Risk of Bias (ROB) assessment tool for the randomized clinical trial, clearly described in (chapter 8.5) of the Cochrane handbook of systematic reviews of interventions 5.1.0. 17,18 The Cochrane risk of bias assessment tool includes the following domain: sequence generation (selection bias), allocation sequence concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selection outcome reporting (reporting bias) and other potential sources of bias. The authors' judgment is categorized as ''low risk'', ''high risk'' or ''unclear risk'' of bias.

Data synthesis
We calculated the Mean Difference (MD) and 95% Confidence Intervals (95% CI), if outcomes were measured in the same way between trials. Standardized Mean Difference (SMD) was calculated in the case of difference in the scoring system.
In some trials, we extracted the data from graphs using PlotDigitizer software. 19 In the case of missing Standard Deviations (SD), we used the available p-values and the calculator available in the Revman software to determine the SD; if unable to calculate, the highest reported SD from the most similar trial for each outcome was used.
Gurlek et al. 2006 2 compared three different corticosteroids (betamethasone, dexamethasone and methylprednisolone) with placebo. For such a study, data related to each steroid versus placebo were included in the metaanalysis as a separate entry designated by the name of the study followed by a capital letter (A, B, or C).
We tested for heterogeneity among included studies by the Chi-Square tests and quantified its extent by the I-Square test. In the absence of clinical and statistical heterogeneity (I 2 less than 50%), the fixed-effect model applied to pool data. In the presence of statistical heterogeneity (I 2 > 50%), the random-effects model for meta-analysis was applied.
All statistical analysis was performed using the RevMan (version 5.3) for Windows software.

Subgroup analysis
We performed subgroup analysis acceding to the types and doses of corticosteroid preparation, in addition to a subgroup analysis of per or postoperative corticosteroid.

Results of risk of bias assessment
We reported an overall low risk of bias according to ROB assessment tool. A summary of risk of bias assessment was reported in Fig. 2. Most of the included studies did not have any clear information about the random sequence generation and the patients' allocation. All of the included studies were double-blinded studies except one, which was a tripleblinded study. 30 All the included studies reported edema and ecchymosis at different days postoperatively.

Qualitative data analysis
We summarized the baseline characteristics of 19 included studies in Table 1. Of them, 18 studies were prospective randomized controlled studies, and only one study was a retrospective clinical trial. 16 A total of 843 patients were enrolled in this systematic review with an average of 30---60 patients per study. We identified a significant difference in the type of corticosteroid used regarding the type and dose.  27 Gurlek et al. 2006 2 was the only study that reported the effect of betamethasone 8 mg for reducing the edema and ecchymosis in rhinoplasty. They compared equivalent doses of three different corticosteroids (dexamethasone 8 mg, betamethasone 8 mg, and methylprednisolone 40 mg) with a placebo. In 2009, Gurlek and colleagues 3 investigated the use of a higher dose of corticosteroids (methylprednisolone 250 mg, 500 mg, and 5000 mg).       NR, not reported. Four studies investigated the methylprednisolone with different doses. Eleven studies investigated the efficacy of combined pre-and postoperative corticosteroid, while eight studies reported preoperative corticosteroid only. Fifteen studies reported the results after rhinoplastic surgery, two studies after orthognathic surgery, 24,25 one study after bilateral sagittal split osteotomies, 22 and one studiy after facelift surgery. 20 The summary of the findings of each included study was reported in the same Table 1.

Quantitative data analysis
A total of 439 patients from 10 clinical trials were included in the meta-analysis. All of these studies reported the assessment of upper and lower eyelid ecchymosis and edema postoperatively using the four-point scale. The most commonly used corticosteroid doses were 8 mg and 10 mg. The higher doses were pooled in one group (corticosteroid > 50 mg). We pooled the results of postoperative ecchymosis and edema at day 1, 3, 5, 7 and 10. The detailed results of subgroup analysis according to the types and doses of the used corticosteroid as well as a subgroup analysis of pre and postoperative corticosteroids are presented in Supplementary Tables 1 and 2.
Tuncel et al. 2013 reported a significant difference between corticosteroid and placebo at day 10 postoperatively.

Unspecified edema
On the first day, the overall effect size of the edema score revealed that edema was significantly reduced in corticosteroids group compared to placebo (SMD = -0.82, 95% CI: -1.37, -0.26). This was not achieved in corticosteroid 8 mg group, but achieved in higher doses as presented in Fig. 7.

Discussion
The present systematic review included 19 clinical trials. Of them, only 10 studies were included in the quantitative evidence synthesis. Our meta-analysis provides Class 1 evidence that corticosteroid has a beneficial effect in reducing postoperative edema and ecchymosis. In addition, preoperative corticosteroids decreased intraoperative bleeding. Although postoperative edema and ecchymosis have commonly occurred after facial plastic surgery in which steroids are commonly used, small numbers of studies were conducted to establish its effectiveness and determine the preferred dose and type to be administered.
On the first postoperative day, smaller doses of corticosteroid (8 mg and 10 mg) were associated with smaller differences in the mean score of upper and lower eyelid edema and ecchymosis, while the higher doses were associated with greater differences. None of the included studies reported any adverse events related to the higher or lower doses of corticosteroid.
Furthermore, preoperative corticosteroid administration significantly reduced the intraoperative bleeding when compared to placebo for higher doses > 50 mg per day (p < 0.0001), but not for 8mg corticosteroid (p = 0.06). Adding postoperative steroid dose to the preoperative one was associated with a decrease in edema and ecchymosis than when given preoperatively alone. Using higher doses of methylprednisolone could improve the postoperative edema and ecchymosis efficiently at day one postoperatively rather than any other steroid. Using a higher methylprednisolone dose was not associated with any intraoperative or postoperative adverse events, but it significantly decreased the intraoperative bleeding. 3,4 None of the included studies reported the results of postoperative pain, healing duration, patient satisfaction, or the quality of life. Corticosteroid has a significant effect in reducing pain following facial plastic, dentoalveolar, and maxillofacial surgeries. 31 In the present meta-analysis, the included studies used the same scale of 0---4 for evaluating the degree of edema and ecchymosis with minimal differences in the description of each grade; therefore, we used the standardized mean difference for each comparison.
Five studies were excluded from the meta-analysis due to their use of different radiological methods, but they reported important results. Abukawa et al. 2017 22 used the CT measurements in masseter muscle thickness and buccal soft tissue to assess the postoperative edema and revealed that the rate of increase in the masseter muscle thickness in the corticosteroid group was significantly lower than that in the control group. On the other hand, Berinstein et al. 1998 15 concluded that the rhinoplasty patients who received dexamethasone had increased postoperative edema when compared to controls as assessed by magnetic resonance imaging scans that quantified the difference in soft tissue thickness between the pre-and postoperative scans.
Using 3D surface scans, Semper-Hogg et al. 2017 24 showed that facial edema significantly decreased postoperatively in the dexamethasone 40 mg group compared to the control group, while lower doses of dexamethasone (5 and 15 mg) did not reduce the facial swelling after orthognathic surgery. Weber and colleagues 28 quantified the facial edema by computer scanning of five sets of standardized photographs and revealed that dexamethasone (8 and 16 mg) administration significantly reduced the postoperative edema.
Alajami and colleagues 23 reported a different method of assessment: the absence of edema was graded as 0, edema of lower eyelid alone was Graded 1, edema of lower and upper eyelids was Graded 2 and edema all around the orbit spreading to the face ± subconjunctival ecchymosis was Graded 3. The result showed that there was highly significant difference in the presence of different grades of edema in the corticosteroid group in all assessment days (p < 0.001). Grade 1 edema was present in the corticosteroid group at day 1 (43.2%), day 2 (29.5%), day 5 (13.6%) and was completely absent from 7th postoperative day onward. In the placebo group, on the 7th postoperative day, 33.3% of the patients still had Grade 1 and 16.7% had Grade 2 edema. Moreover, 16.7% in the placebo group continued to show Grade 1 edema on the 10th postoperative day.
Regarding the intensity of ecchymosis and severity of edema, Totonchi et al. 2007 26 showed that dexamethasone 10 mg reduced edema during the early postoperative period with no significant differences in the ratings of extent and intensity of ecchymosis on postoperative day 2, but there was a significant difference for the edema rating, with the control group demonstrating more swelling compared with dexamethasone group. On postoperative day 8, the dexamethasone group demonstrated a larger extent and higher intensity of ecchymosis compared with the control group and there were no differences in the magnitude of edema by postoperative Day 8.
Tuncel et al. 2013 16 reported that dexamethasone with controlled hypotension considerably reduced intraoperative bleeding, postoperative swelling and ecchymosis of rhinoplasty. Including this study in our meta-analysis was associated with significant heterogeneity because it reported the comparative data as median and interquartile range, which required farther data transformation to be pooled with other studies. Removing Tunnel's study from the analysis led to resolving the heterogeneity and was not associated with the change in the overall significance in all outcomes.
The overall quality of the evidence of the included studies is high. The main concern is the small sample size of each article. According to the American Society of Plastic Surgeons report 2018, Nose reshaping is one of 2018 s top 5 cosmetic surgical procedures (213,000 surgeries in 2018). 32 Although we are presenting the first large comprehensive systematic review (19 included studies) to investigate the effect of corticosteroids on reducing the post-facial plastic surgery complications, we reported a consistent result with the previous Cochran review in 2014 33 and with other systematic reviews that investigated the use of steroids in rhinoplasty. 34,35 The Cochran review reported that a single preoperative dose of 10 mg dexamethasone decreased edema and ecchymosis over the first two postoperative days; this difference was not observed after this period, while high doses of methylprednisolone decreased both ecchymosis and edema at days 1, 3 and 7 postoperatively.
Most of the included studies were heterogeneous regarding the numbers and dosage of pre-and postoperative corticosteroids as well as the type of the steroids used.. Studies by Abukawa et al., 22 Griffies et al., 13 Koc et al., 4 and Ozdel et al. 21 described giving steroid preoperatively while studies by Tuncel et al., 16 Mehdizadeh et al., 30 Kargi et al., 29 Hoffmann et al., 14 and Alajami, 23 gave steroids as both a pre-and postoperative dose. Therefore, we performed subgroup analysis to pool the similar groups together in one meta-analysis model and provide an overall effect estimate as seen in supplementary Tables 1---2. Most of the subgroups were underpowered and the result cannot be generalized due to the small numbers of included studies. We recommend conducting well-designed randomized controlled trials with a large sample size to investigate the effect of timing (pre and postoperative) and dosage (single and multiple) and type of corticosteroid given for facial plastic surgery.
In conclusion, this comprehensive meta-analysis confirms a statistically significant benefit of the preoperative administration of corticosteroid when compared to placebo. Furthermore, continuing the steroids postoperatively is associated with long-term reduction of complications, including edema, ecchymosis, and intraoperative bleeding. Finally, higher doses of corticosteroids are associated with a more significant reduction in edema and ecchymosis, but further studies are recommended to determine the postoperative side effects such as surgical site infection and delayed healing.

Conflicts of interest
The authors declare no conflicts of interest.