FLEX January 2024

10976817, 2019, 5, Downloaded from https://aao-hnsfjournals.onlinelibrary.wiley.com/doi/10.1177/0194599818823205 by National Institutes Of Health, Wiley Online Library on [12/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

752

Otolaryngology–Head and Neck Surgery 160(5)

syndromes affecting the IAC as the cause of PT. 19 We excluded studies that used endovascular procedures, such as embolization, venous ligation, and straightening of tortuous vessels. We also excluded nonresurfacing extravascular pro cedures, such as intracranial vascular decompressions and compression or obliteration of the sigmoid sinus, as one of the goals during resurfacing is to avoid compressing the sig moid sinus to mitigate the risk of increased intracranial pressure. We excluded studies that selectively reported suc cessful cases. Two authors (G.S.L. and B.C.B.) assessed the eligibility of studies, with any disagreements settled by consensus with a third author (Y.V.). Initial screening was done with the article title and abstract to generate a list of potentially rele vant studies. Because databases overlapped in their coverage of records, duplicate studies were removed with Zotero soft ware prior to initial screening. Each study in the list then underwent full-text review to determine eligibility by our inclusion criteria. If different publications reported on sub sets of duplicative data, we included the study with the larg est amount of data (eg, the study that followed the same cohort for the longest period). For non-English articles, the full text was assessed by a medical expert fluent in the lan guage or translated into English with Google Translate. Data Extraction For each included study, we extracted data about the partici pants’ characteristics (eg, country, average age, sex), sided ness of PT, data collection period, diagnosed vascular etiology of PT, preoperative imaging, surgical approach, material used for resurfacing or sound blockade, mean length of follow-up, postoperative resolution of PT, pre- to postoperative audiometric changes, and complications. Postsurgical outcomes were based on patients’ symptoms of PT at last follow-up and were classified as follows: com plete (95%-100% resolution of PT), partial (1%-95% resolu tion of PT or complete resolution of PT with eventual recurrence of PT not better explained by another trigger), no resolution (0% change in PT), or worsened. Subjective and questionnaire-based measures of PT severity were both used to classify PT postsurgical outcomes. Materials used for resurfacing were classified by source as autologous (temporalis fascia or muscle, tragal cartilage, bone chips, bone dust) or artificial (bone cement, xenograft). We also classified materials by density as hard (hydroxya patite cement, bone, bone dust) or soft (cartilage, perichon drium, muscle, fascia). Bone wax and dissolvable materials such as Surgicel (Ethicon), Gelfoam (Pfizer), and fibrin glue were not considered permanent resurfacing materials for classification and were excluded from analysis. Assessment of Bias We assessed risk of bias in studies with the guidelines set forth in the Cochrane Assessment Tool. 20 Each included study was evaluated according to 2 domains: selective out come reporting and incomplete outcome data. For each domain, a study was assessed as low or high risk of bias.

This is similar to the approach used in another systematic review. 21 For selective outcome reporting, a study was considered low risk if inclusion/exclusion criteria were described and all of the following postoperative outcomes were reported: tinnitus change, hearing change (eg, by audiogram), and presence of any complications. A study was considered high risk if at least 1 of these expected postoperative outcomes was not reported or if inclusion/exclusion criteria were not described. For incomplete outcome data, a study was considered low risk if all outcomes were reported consistently for all patient cases in the study. A study was considered high risk if certain outcomes were reported for some patients but not others. We assessed risk of incomplete outcome data only for studies involving . 1 resurfacing case. Statistical Analysis We performed descriptive statistics with Google Sheets cloud software. For categorical variables, P values were obtained by Pearson’s chi-square test with Microsoft Excel 2013. A critical value of .05 was used to assess significance with P values. Results Our search terms produced 400 results on PubMed, 539 results on Embase, and 16 controlled trials on Cochrane Database, for a total of 954 records. After removal of dupli cate studies, 571 records remained. Hand search of reference lists in articles identified an additional 5 publica tions. 8,14,15,22,23 We performed initial screening of these 576 records by titles and abstracts, yielding 72 that were assessed by full-text review for eligibility. Among these, 20 studies met inclusion criteria and were included 4-10,14,22,24-34 ( Figure 4 ). Five publications 15,35-38 were excluded due to overlapping cohorts with later follow-up studies. Three of these studies 15,35,36 provided information on cohort mem bers documented in a later and more inclusive publication by Eisenman et al 4 ; 1 publication 37 was more inclusively documented by Wang et al 27 ; and 1 publication 38 was more inclusively documented by Zeng et al. 24 Additionally, 2 studies were excluded because their full texts were not available. This review included 20 studies published between 1985 and 2018 and incorporated data collected from 9 countries (Belgium, Australia, Egypt, China, Korea, Japan, Singapore, United States, and Canada; Table 1 ). Thirteen studies reported about sigmoid sinus dehiscence or diverticulum, 5 about jugular bulb dehiscence or diverticulum, and 5 about dehiscent or aberrant ICA. All full texts that were included were published in English except for 2 that were in Chinese. 32,33 All except 1 of the included studies incorporated preo perative computed tomography (CT) imaging to assess for major vessel wall anomaly among patients presenting with a chief complaint of PT. Rouillard et al 31 relied on otoscopy to identify a bluish hue on the posteroinferior quadrant of