Prevalence of Frey syndrome following extraoral surgical treatment for mandibular fractures: a systematic review and meta-analysis

Search strategy

The Medline (PubMed search engine) and Scopus database searches were conducted in a thorough manner, adhering to the guidelines set by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA),¹⁶ ensuring a meticulous and rigorous approach as depicted in Figure 1. The PRISMA checklist was utilized to facilitate the systematic review process. Articles published up until May 1st, 2023, were collected. An independent literature search was conducted by two reviewers using the following keywords: “mandibular”, “mandible”, “jaw”, “fractures”, “open reduction and internal fixation”, “ORIF”, “surgical treatment”, “Frey syndrome”, “Baillarger’s syndrome”, “gustatory hyperhidrosis”, “auriculotemporal syndrome”, “Dupuy syndrome”, “prevalence”, “incidence”, “rate”. In conjunction with the primary search, a thorough examination of the reference lists from the identified studies was conducted to identify any additional articles that may have been overlooked. The collected studies were meticulously organized and stored utilizing the Zotero reference management software (version 6.0.26).¹⁷ The credibility of our dataset was ensured by diligently removing any duplicate references. Following the initial search, the remaining articles were thoroughly examined by two independent investigators. The study selection process comprised two distinct stages. Initially, the titles and abstracts of the articles were meticulously reviewed, leading to the elimination of those not meeting the predetermined inclusion criteria. In the second stage, the full texts of the remaining articles were obtained and subjected to a comprehensive evaluation. Any disagreements during the study selection were resolved through consensus among the team members, ensuring a consistent and unified decision-making process. By employing this systematic approach, a comprehensive and dependable collection of studies for analysis was achieved.

Figure 1. Flow chart illustrating the organized presentation of outcomes derived from the identification and choice of pertinent research studies.

Criteria for study selection and data extraction

Articles that specifically investigated the prevalence of Frey syndrome following surgical interventions through extraoral incision for mandibular fractures were included with no restriction on publication date. Excluded from consideration were case reports, case series involving fewer than five participants, review papers, systematic reviews, meta-analysis, randomized or non-randomized clinical trials, studies involving animals, studies lacking complete text, articles not in English language, studies regarding other approaches (intraoral),¹⁸ articles examining Frey syndrome per fracture and articles containing data derived from surveillance databases. The subsequent factors were extracted from each study: the name of the primary author, publication year, research design, geographical region of origin, country of origin, study duration, overall number of patients, percentage of male participants, mean age, and individuals experiencing postoperative Frey syndrome.

Quality assessment

To evaluate the quality of the studies included, two investigators independently assessed them using the National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment tool for Observational Cohort and Cross-Sectional Studies. The evaluation process entailed a thorough examination of each study to identify any methodological or survey implementation weaknesses that could impact internal validity. During the assessment, the investigators considered 14 specific questions to gauge the quality of each study. They were provided with response options such as “yes,” “no,” “cannot determine” (e.g., in instances where the data presented uncertainties or contradictions), “not reported” (e.g., in cases where data were not reported or were incomplete), or “not applicable” (e.g., when a question did not pertain to the specific type of study under evaluation). By evaluating these questions, the investigators categorized the risk of bias for each study as either “low,” “moderate,” or “high,” enabling an overall assessment of the study’s quality.¹⁹ By conducting this rigorous quality appraisal, our aim was to ensure that only studies demonstrating a moderate or high level of internal validity were included in our analysis.

Statistical analysis

Using RStudio software (version: 2022.12.0+353) by RStudio Team (2022),²⁰ a meta-analysis was executed. The meta-analysis, facilitated by the metafor package,²¹ involved estimating the pooled prevalence and its associated 95% confidence intervals (CI) using the DerSimonian and Laird random-effects model, complemented by a Freeman-Tukey double arcsine transformation.²² The existence of heterogeneity among studies was assessed by visually examining the forest plot and employing Cochran’s Q statistic along with its corresponding p-value. The Higgins I² statistic, accompanied by its 95% CI, was employed to quantify the extent of genuine heterogeneity in effect sizes. An I² value falling within the ranges of 0%-40%, 30%-60%, 50%-90%, and 75%-100% denoted insignificant, moderate, substantial and considerable heterogeneity, respectively. To determine if the potential outlying effect sizes were also influential, screening for externally studentized residuals with z-values larger than two in absolute value and leave-one-out diagnostics were performed.²³ Due to a paucity of data regarding categorical variables, such as surgeon’s level, subgroup analysis was not performed. In the conducted meta-regression analysis involving continuous variables, an investigation was carried out to assess the influence of publication year and the proportion of male participants as moderators impacting the relevant effect sizes. Due to the limited availability of data on additional continuous variables like the mean age of the patients, the analysis was performed using the aforementioned ones.²⁴ Unless otherwise stipulated, the statistical significance was established at p = 0.05 (two-tailed). Methods designed to assess publication bias, including Egger’s test²⁵ and Begg’s test²⁶ along with the utilization of funnel plots, were originally formulated within the framework of comparative data. These methods operate under the assumption that studies with positive outcomes are more likely to be published than those with negative outcomes. Nevertheless, in a meta-analysis centered on proportions, the criteria for defining a positive result lacks clarity and consensus.²⁷ Consequently, in this current meta-analysis, an evaluative approach was adopted to qualitatively assess potential publication bias.

Results and characteristics of the included studies

In aggregate, this analysis encompassed a grand total of 15 studies, collectively involving a cohort of 622 participants. These publications spanned the period from 2000 to 2020, encapsulating research carried out between 1990 and 2019. Among these studies, one adopted a cohort design,⁴² while the remaining were cross-sectional in nature.^28‐41 Geographically, the studies were primarily executed across Asia (including India, Japan, Turkey, and Nepal), followed by Europe (including Italy, France, Austria, and Switzerland), and finally America (including the USA and Canada). The representation of male participants averaged at 79.9%, and the mean age of participants ranged from 27 to 41.5 years, with a median age of 29.2 years. The assessment of study quality consistently categorized all the studies as being of moderate quality. Their detailed characteristics are outlined in Table 1.

Prevalence of Frey syndrome following extraoral surgical treatment for mandibular fractures

A random-effects model analysis yielded an initial overall Frey syndrome prevalence following ORIF of 0.01% (95%CI 0%-0.59%) with moderate between studies heterogeneity I²=0% (95%CI 0%-44%, p=0.84). The influence diagnostics and the forest plot illustrating the results of the leave-one-out analysis is presented in Figures 2 and 3. As per them, the study conducted by Zrounba et al.³⁸ identified as influential. After the exclusion of the aforementioned study the estimated prevalence was calculated at 0.01% (95%CI 0%-0.7%) with moderate remaining between-studies heterogeneity I²=0% (95%CI 0%-47%) (p=0.79) (Figure 4).

Figure 2. Visual representation of the influence diagnostics for each of the included studies regarding the prevalence of Frey syndrome after ORIF for mandibular fractures.

Figure 3. Forest plot displaying the re-calculated pooled effects, with one study omitted each time, using the leave-one-out method.

Figure 4. Forest plot evaluating the calculated prevalence of Frey syndrome following extraoral surgical treatment for mandibular fractures.

Meta-regression analysis

The meta-regression analysis, which employed continuous variables, unveiled that there exists no statistically significant relationship, whether positive or negative, between the year of publication and the proportion of males. This outcome is visually represented in the Table 2.

Study’s strengths and limitations

The principal strength of the present study is grounded in the methodological approach employed throughout its entirety. However, it is essential to acknowledge the presence of several limitations within this study. Notably, the moderate level of unaccounted heterogeneity present among the included studies does introduce a degree of caution when interpreting the results. As previously noted, the intrinsic nature of this specific study category inherently gives rise to expected outcome variations, thus contributing to the observed moderate heterogeneity. Additionally, irrespective of study nature, potential confounding factors can introduce bias to estimated Frey syndrome prevalence after ORIF for mandibular fractures. These factors (e.g., extended operative durations, patient age, additional concurrent procedures) underline the need for nuanced consideration of the study’s findings. An attempt was made to conduct a subgroup analysis involving variables such as the localization of mandibular fractures or the specific type of extraoral approaches used, with the intention of enhancing the study’s depth. Practical challenges arising from the intricacies of information within each study necessitated a thorough evaluation of potential trade-offs. Opting for a broader scope with the available data was deemed imperative to maintain the overall integrity and statistical robustness of the study. This decision was made to prevent compromising the reliability and statistical power of the study, recognizing that excluding these studies could have had adverse effects.

Furthermore, the inclusion of solely English-language observational studies led to the emergence of reporting bias.