Introduction

In the modern era of rhinoplasty, the surgeon recognized the importance of achieving an improved quality of life outcome in contradistinction to merely a surgical result and it starts by properly selecting patients (Adamson & Gantous, 2019; Tasman, 2010). Formal evaluation by a psychiatrist of candidates for rhinoplasty has already shown a noticeable prevalence of psychiatric comorbidities and the major psychiatric diagnosis was body dysmorphic disorder (BDD; Alavi et al., 2011; Belli et al., 2013). Despite many cosmetic professionals report knowledge of diagnostic criteria and clinical presentation of BDD, they hardly identify any patients with these disorders (Kattan et al., 2020).

The prevalence of body dysmorphic disorder in the general population is around 1.9% and jumps to 11.3% in dermatological populations, 13.2% in general cosmetic surgery patients, and 20.1% in rhinoplasty patients (Veale et al., 2016). People with BDD are concerned about flaws in their physical appearance that are not observable or appear slight to others but cause them emotional and social distress and lead to repetitive behaviors or thoughts. BDD is classified under the obsessive–compulsive disorder (OCD) and related conditions chapter and, while some patients have good insight, others are completely delusional (APA, 2022; WHO, 2019). As a result, these individuals tend to pursue cosmetic treatment, but this usually does not improve their dysmorphic symptoms and they often request repeated surgeries (Phillips, 2005).

Although rhinoplasty is known to be the surgery most frequently sought out by people with BDD (De Brito et al., 2016; Dey et al., 2015; Phillips, 2005), the prevalence of this disorder in patients seeking for this procedure falls within 1.8% to 59% (De Brito et al., 2015; Picavet et al., 2013). Probably, this wide range makes screening prior to procedures less of a priority. Additionally, the nose is among the five most requested body parts for correction through plastic surgery in the world and the population of rhinoplasty candidates continues to rise (ISAPS, 2019). To further investigate this variation, we conducted a systematic review and meta-analysis of studies that addressed the prevalence of BDD or dysmorphic symptoms in rhinoplasty candidates from otorhinolaryngology or plastic surgery clinics.

Methods

Search Strategy and Study Eligibility

Cross-sectional and longitudinal studies published up to April 2021 that reported on the prevalence of BDD or dysmorphic symptoms in rhinoplasty patients from otorhinolaryngology (ear-nose-throat, ENT) or plastic surgery clinics were identified using PubMed, Cochrane, and SciELO (independently performed by the first two authors); and by screening the reference lists of articles identified using the approach recommended by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021). The computer-based searches combined MeSH terms related to rhinoplasty (“Rhinoplasty”; “Plastic Surgery”; “Esthetic Surgery”; “Cosmetic Surgery”), with those related to BDD (“Body dysmorphic disorder”; “Somatoform disorder”; “Body image disorder”; “Body image disfunction”; “Dysmorphophobia”). There was no language restriction for the inclusion of full texts. The review protocol was registered in PROSPERO before data collection (full details of the search strategy and the registration of the protocol are presented in Online Resource 1). Studies were included if they used a validated method to assess for BDD or dysmorphic symptoms. The impossibility of acquiring the full text online was a reason for exclusion. After assessing possible biases, eligible studies were selected for meta-analysis.

Data Extraction and Quality Assessment

The following information was extracted from each article using a standardized form: authors, publication year, geographic location, enrollment period, medical specialty sought, inclusion and exclusion criteria (including reason for rhinoplasty and previous rhinoplasty), diagnostic or screening method used, sample size, number and percentage of female participants, age, number of BDD positive participants, and estimated prevalence of BDD or dysmorphic symptoms. The most comprehensive publication was used when several studies involved the same population of rhinoplasty patients.

The Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Studies Reporting Prevalence Data (Munn et al., 2015) was used to assess the quality of all studies included in the systematic review by evaluating the presence of appropriate sample frame and size, detailed description of study subjects and settings, adequate statistical analysis, etc. According to the percentage of yes responses, studies were judged to be at low risk of bias (≥ 70%) or high risk of bias (< 50%). All discrepancies were resolved by discussion and opinions of the other two reviewers. This instrument was used as merely a descriptive measure and was not an exclusion criteria.

Data Synthesis and Analysis

The prevalence of BDD in each sample was calculated by dividing the number of diagnosed/screened cases of BDD by the total number of rhinoplasty patients. Binomial proportion confidence intervals for individual studies were calculated using the Clopper-Pearson method. The main characteristics of the studies included in the qualitative synthesis were tabulated and Forest plots were created for the quantitative synthesis. To pool the prevalence data we used random-effects meta-analysis. Between-study heterogeneity was assessed by standard Cochran’s χ2 tests (Q-test) and the I2 statistic and by comparing results from studies grouped according to prespecified study-level characteristics (geographic location, specialty sought, reason for rhinoplasty, only primary rhinoplasty seekers, and diagnostic method) using stratified meta-analysis. Variability between studies (τ2) was calculated with the DerSimonian-Laird estimation method. The weighting of each study was performed using the inverse variance method. Publication bias was assessed via visual inspection of the funnel plot and Begg test. All analyses were performed using Microsoft Excel 365 and RStudio Version 1.4.1717 software, with emphasis on the Meta library version 3.2–0. Statistical tests used a significance threshold of p < 0.05.

Results

Study Characteristics

Twenty-one studies were eligible for qualitative synthesis (Table 1). Twelve studies involving a total of 1318 individuals were included in the meta-analysis (Alavi et al., 2011; Baykal et al., 2015; Belli et al., 2013; De Brito et al., 2016; De Souza et al., 2021; Joseph et al., 2017; Lekakis et al., 2016; Rabaioli et al., 2020; Saeed et al., 2021; Spataro et al., 2020; Strazdins et al., 2017; Veale et al., 2003; Fig. 1 and Fig. 2). Four studies were conducted in Asia (Alavi et al., 2011; Baykal et al., 2015; Belli et al., 2013; Saeed et al., 2021), three in Europe (Joseph et al., 2017; Lekakis et al., 2016; Veale et al., 2003), three in South America (De Brito et al., 2016; De Souza et al., 2021; Rabaioli et al., 2020), one in North America (Spataro et al., 2020), and one in Oceania (Strazdins et al., 2017).

Table 1 Selected characteristics of the 21 studies included in this systematic review
Full size table
Fig. 1
figure 1

Flow diagram depicting the process of identifying studies on the prevalence of body dysmorphic disorder and dysmorphic symptoms in rhinoplasty patients. BDD, body dysmorphic disorder

Full size image
Fig. 2
figure 2

Meta-analysis of prevalence of body dysmorphic disorder and dysmorphic symptoms in rhinoplasty candidates. n.e, number of events; n, number of participants; CI, confidence interval

Full size image

The informed recruitment period ranged from 2007 to 2019 and the sample size ranged from 29 to 306. Veale et al. (2003) and Belli et al. (2013) did not inform the period of data collection. The participants were mostly female with a mean age ranging from 22.8 to 38 years. Few studies characterized data on marital status with a predominance of singles in the majority (Belli et al., 2013; De Brito et al., 2016; De Souza et al., 2021; Lekakis et al., 2016; Saeed et al., 2021; Spataro et al., 2020), but married individuals predominated in the study with the largest sample size (Alavi et al., 2011).

Eight (67%) studies recruited rhinoplasty patients from only ENT clinics (Baykal et al., 2015; Belli et al., 2013; De Souza et al., 2021; Joseph et al., 2017; Lekakis et al., 2016; Rabaioli et al., 2020; Spataro et al., 2020; Strazdins et al., 2017). Four studies (Alavi et al., 2011; Belli et al., 2013; Lekakis et al., 2016; Veale et al., 2003) included patients who seeked rhinoplasty only for esthetic reason, five (De Souza et al., 2021; Joseph et al., 2017; Rabaioli et al., 2020; Spataro et al., 2020; Strazdins et al., 2017) included esthetic and/or functional motivation and three (Baykal et al., 2015; De Brito et al., 2016; Saeed et al., 2021) did not informed the reason for rhinoplasty.

Three studies (Baykal et al., 2015; Joseph et al., 2017; Veale et al., 2003) assessed for BDD symptoms using the Body Dysmorphic Disorder Questionnaire (BDDQ; Phillips, 2005), two (Lekakis et al., 2016; Spataro et al., 2020) used the BDDQ-Aesthetic Surgery Version (BDDQ-AS; Lekakis et al., 2016), two (Saeed et al., 2021; Strazdins et al., 2017) used the Dysmorphic Concern Questionnaire (DCQ; Oosthuizen et al., 1998), and one (De Souza et al., 2021) used the Body Dysmorphic Symptom Scale (BDSS; Perugi et al., 1997). Two (Alavi et al., 2011; Belli et al., 2013) diagnosed BDD using structured psychiatric interviews (APA, 1994; First et al., 2002), and two (De Brito et al., 2016; Rabaioli et al., 2020) used the interview version of the Body Dysmorphic Disorder Examination (BDDE; Rosen & Reiter, 1996).

When evaluated according to the JBI Critical Appraisal Checklist, 90% of the 21 studies did not attain a positive score of 70%, and 48% did not even score 50% (see Table, Online Resource 2, which lists the quality score obtained by each study). Nine articles were not included in the meta-analysis because of significant methodological biases. Selection bias was considered in De Brito et al. (2015) and in Ramos et al. (2016), who pre-selected and only evaluated for BDD rhinoplasty candidates showing excessive concern with physical appearance associated with clinically significant subjective distress and in Jeremy and Stephen (2014) and Bulut et al. (2018), studies that only addressed postoperative rhinoplasty patients. A study by Milad et al. (2019) was also excluded because it was not possible to determine whether only rhinoplasty candidates were included. Measurement bias was considered in Ghadakzadeh et al. (2011), Crerand (2003), Picavet et al. (2011), and Picavet et al. (2013), since surgeons graded the nasal deformity and only considered participants with minimal or absent deformity as positive for BDD.

Prevalence of Body Dysmorphic Disorder and Dysmorphic Symptoms in Rhinoplasty Patients

Meta-analytic pooling of the prevalence estimates for BDD or dysmorphic symptoms reported by the 12 studies yielded a summary prevalence of 34.1% (449/1318, 95% CI, 27.2–42.1%), with significant evidence of between-study heterogeneity (p < 0.01, τ2 = 0.28, I2 = 86%) (Fig. 2). Despite the high heterogeneity found, we chose not conducting sensitivity analysis due to the lack of outliers and we preferred the analysis by study-level characteristics to overcome this.

Prevalence of Body Dysmorphic Disorder and Dysmorphic Symptoms by Study-Level Characteristics

The prevalence of BDD or dysmorphic symptoms significantly decreased when comparing studies whose participants were only primary rhinoplasty candidates (Alavi et al., 2011; Belli et al., 2013; Veale et al., 2003) and those that included candidates for primary or revision rhinoplasty (Baykal et al., 2015; Lekakis et al., 2016; Spataro et al., 2020; Strazdins et al., 2017; 87/385, 23% [CI 95%, 19.1–27.2%] vs 156/450, 38.2% [95% CI, 26.4–52.3%]; test for subgroup differences, Q = 5.71, p = 0.017). On the other hand, higher prevalence was observed in studies from plastic surgery (De Brito et al., 2016; Saeed et al., 2021) vs otorhinolaryngology clinics (Baykal et al., 2015; Belli et al., 2013; De Souza et al., 2021; Joseph et al., 2017; Lekakis et al., 2016; Rabaioli et al., 2020; Spataro et al., 2020; Strazdins et al., 2017; 125/261, 48% [CI 95%, 40.2–55, 4%] vs 243/722, 33.3% [CI 95%, 25.3–42.9%]; Q = 5.34, p = 0.021).

Among the full set of studies, no statistically significant differences in prevalence estimates were noted between studies conducted in eastern world (Alavi et al., 2011; Baykal et al., 2015; Belli et al., 2013; Saeed et al., 2021) and western world (De Brito et al., 2016; De Souza et al., 2021; Joseph et al., 2017; Lekakis et al., 2016; Rabaioli et al., 2020; Spataro et al., 2020; Strazdins et al., 2017; Veale et al., 2003; 158/522, 31.2% [95% CI, 17–50.1%] vs 291/796, 35.3% [95% CI, 28.2–42.9%]; Q = 0.16, p = 0.685), between studies in Brazil (De Brito et al., 2016; De Souza et al., 2021; Rabaioli et al., 2020) and elsewhere (Alavi et al., 2011; Baykal et al., 2015; Belli et al., 2013; Joseph et al., 2017; Lekakis et al., 2016; Saeed et al., 2021; Spataro et al., 2020; Strazdins et al., 2017; Veale et al., 2003; 149/339, 42.9% [95% CI, 33.8–52.1%] vs 300/979, 31.2% [95% CI, 23.4–40.3%]; Q = 3.21, p = 0.073), between studies that only addressed esthetic motivations for seek rhinoplasty (Alavi et al., 2011; Belli et al., 2013; Lekakis et al., 2016; Veale et al., 2003) and those that considered esthetic and/or functional reasons (De Souza et al., 2021; Joseph et al., 2017; Rabaioli et al., 2020; Spataro et al., 2020; Strazdins et al., 2017; 142/501, 25.1% [95% CI, 14.4–40%] vs 151/500, 31.4% [95% CI, 25.2–38.2%]; Q = 0.99, p = 0.32), and between studies that assessed BDD via self-applied screening tools (Baykal et al., 2015; De Souza et al., 2021; Joseph et al., 2017; Lekakis et al., 2016; Saeed et al., 2021; Spataro et al., 2020; Strazdins et al., 2017; Veale et al., 2003) and those that applied diagnostic interviews (Alavi et al., 2011; Belli et al., 2013; De Brito et al., 2016; Rabaioli et al., 2020; 239/680, 36.1% [95% CI, 27.3–44.8%] vs 210/638, 30.3% [95% CI, 18.1–47.3%]; Q = 0.34, p = 0.562; Fig. 3).

Fig. 3
figure 3

Meta-analysis of prevalence of body dysmorphic disorder and dysmorphic symptoms in rhinoplasty candidates stratified by study-level characteristics. Legend: the area of each diamond is proportional to the inverse variance of the estimate. No., number; BDD, body dysmorphic disorder; CI, confidence interval

Full size image

Assessment of Publication Bias

Visual inspection of the funnel plot (Fig. 4) and Begg test (p = 0.6384; Kendall’s tau =  − 0.1212) did not reveal significant asymmetry, and the publication bias hypothesis was discarded.

Fig. 4
figure 4

Assessment of publication bias via funnel plot and Begg test. Legend: the linear regression test of funnel plot asymmetry discarded the presence of publication bias (p = 0.6384)

Full size image

Discussion

The meta-analysis of the prevalence of BDD and dysmorphic symptoms in 1318 rhinoplasty preoperative patients found a rate of 34% and variation of 27.2–42.1%. By subgroups, a prevalence rate of 23% was found in candidates for their first rhinoplasty, 48% among patients who sought plastic surgeons for rhinoplasty, and 33% among patients that went to ENTs for this procedure. We obtained a good global representation as our meta-analysis included studies from all continents except Africa. The higher prevalence of women was already expected since the female population seeks rhinoplasty and other cosmetic surgeries or procedures more frequently than men (ISAPS, 2019). The prevalence of BDD did not appear to change over the years.

Our findings confirm the high prevalence of BDD in the rhinoplasty population, but it is even higher than the value obtained in the systematic review by Veale et al. (2016), which found a balanced prevalence of 20.1% in this same population. However, the seven studies included in Veale’s analysis were highly heterogeneous (I2 = 90.5%) and included four samples that were excluded from this present study because of how the patients were selected (Ghadakzadeh et al., 2011), unvalidated/uncertain means of assessing BDD (Constantian, 2012; Fathololoomi et al., 2013), or the surgeon’s use of defect scales (Picavet et al., 2013).

Our decision to exclude those studies that only considered patients positive for BDD if the surgeon determined the nasal deformity was minimal or non-existent (Crerand, 2003; Picavet et al., 2011, 2013) was based on the findings of Picavet et al. (2011) and Picavet et al. (2013), in which the proportion of patients positive for BDD after the surgeon’s assessment of the defect was lower than expected (2.2% and 1.8%, respectively), despite the large proportion of patients with moderate/severe dysmorphic symptoms (33% and 36%, respectively). Crerand (2003) has already demonstrated that surgeons tend to point out more defects as observable than lay evaluators, so the use of defect scales by the surgeon could increase the number of false negatives. Another issue is basing the diagnosis only on evaluation of the nose, when patients with BDD may refer to several parts of their body as flawed (Baykal et al., 2015; De Brito et al., 2016; Phillips, 2005). Another critique to the studies by Picavet et al. (2011) and Picavet et al. (2013) is the use of the BDD-YBOCS to diagnose BDD by establishing a cut-off point, which runs counter to the recommendations for using this scale (Phillips, 2005, 2012). The prevalence data obtained from inappropriate use of the BDDE-SR (Rosen & Reiter, 1995) in the study by Baykal et al. (2015) were also excluded. In this sense, by excluding from the analysis those studies that could contain more false negatives, we should expect an increase in the prevalence of BDD.

The higher prevalence found when we analyzed studies in which participants were seeking their revision rhinoplasty was already anticipated, considering that previous plastic surgery of the nose is a red flag indicating BDD should be considered. Surgical treatment is not usually effective in these patients because their body image is distorted, and as a result they tend to be dissatisfied after surgery and seek repeated procedures (Davis & Bublik, 2012; Phillips, 2005). Picavet et al. (2013) reported more severe BDD symptoms in patients with a prior history of rhinoplasty (p < 0.001).

A significantly higher proportion of moderate to severe BDD symptoms is also predicted in individuals who pursue rhinoplasty for esthetic reasons compared to for functional reasons (Picavet et al., 2011). We assume that traditionally plastic surgeons are sought out by patients seeking exclusively esthetic improvements, while ENT specialists attract more patients who wish to improve nasal function (as a single or an additional motive). As a result, a higher prevalence of BDD was already expected among individuals who seek this procedure via plastic surgeons. However, the vast majority of the studies found during the review were conducted in tertiary centers: the participants were referred, and may not have actively sought out a specific type of specialist for their rhinoplasty. Therefore, the attending physician may have chosen the specialist.

However, our attempt to compare the meta-analyses of prevalence according to the motivation for rhinoplasty yielded inconclusive results, mainly due to the major heterogeneity and variability among the subgroup seeking esthetic surgery. It is also important to point out that the two studies of patients who sought plastic surgeons were not included in this subgroup (De Brito et al., 2016; Saeed et al., 2021), along with the study by Baykal et al. (2015), because these studies did not explain the patients’ motives for rhinoplasty in the methodology section. On the other hand, the studies within the esthetic and/or functional motivation subgroup included different proportions of patients seeking esthetic or functional improvement, or both. In Strazdins et al. (2017) and De Souza et al. (2021), all participants were candidates for esthetic as well as functional surgery. Despite the significant difference in the proportion of esthetic and functional or exclusively functional cases reported by Rabaioli et al. (2020) and Spataro et al. (2020), (74% and 21% versus 32% and 20%, respectively), this does not appear to be reflected in the prevalence rates found (38.2% and 31.6%), but the former study reported significantly increasing severity scores for BDD of rhinoplasty for functional, both, or esthetic motives. Joseph et al. (2017) did not report the proportions of their patients in terms of motivation, but did state that one patient with functional motivation who was previously negative for BDD became BDDQ positive during postoperative follow-up.

The only study in the systematic review that focused exclusively on patients with functional reasons for rhinoplasty was by Bulut et al. (2018); these researchers found a 10% prevalence of BDD measured five years after surgery after applying screening measures (DCQ and BDDQ). However, just as Felix et al. (2014) found 81% remission of BDD 1-year post-rhinoplasty in 31 women with mild to moderate BDD, it is not possible to rule out that the surgical treatment employed influenced the identification of BDD. Furthermore, the surgeon may have avoided operating on patients in cases where potentially unfavorable psychopathological characteristics were identified prior to surgery. For these reasons, this study was not included in the meta-analysis.

On the other hand, a significant difference was predicted between prevalence studies that assessed BDD through diagnostic interviews and those involving self-applied questionnaires for the same purpose, considering that these latter are screening methods and consequently carry a higher risk of false positives (Veale et al., 2003). The high heterogeneity (\({I}^{2}\)= 93%) and variability (\({\tau }^{2}\)= 0.4846) between the evaluated studies likely resulted from the way the diagnosis was made, and may have contributed to these findings. The studies by Alavi et al. (2011) and Belli et al. (2013), which diagnosed BDD based on a semi-structured interview with psychiatrists (the gold standard), found lower prevalence rates than the studies by de Brito et al. (2016) and Rabaioli et al. (2020), which diagnosed BDD through the interview version of the BDDE applied by an experienced psychologist and unspecified trained interviewer, respectively (25% and 12% versus 52% and 38%). Another reason for the major variation in prevalence between these four studies that used different methods to diagnose BDD may be the country of origin, since both studies with higher prevalence rates are Brazilian. But an additional possibility is that in a population where a certain condition is prevalent (which is the case of BDD in rhinoplasty patients), a high positive predictive value of the tests used to identify this condition is also expected, and for this reason, it may be that the prevalence rates found from the screening tests are close to those found using diagnostic methods (Kawamura, 2002).

The meta-analysis that included only Brazilian studies yielded 43% prevalence, which tended to be higher than the rate found in the meta-analysis of the other countries (31%). Global statistics indicate that rhinoplasty and plastic surgeries in general are extremely popular in Brazil (ISAPS, 2019). On the other hand, it has also been assumed that the prevalence of BDD would be higher in predominantly Islamic countries where traditional women’s clothing leaves only the face exposed and rhinoplasty is a popular surgery among this population (Baykal et al., 2015; Ghadakzadeh et al., 2011). However, we found no difference in prevalence between Eastern and Western world. The countries representing the eastern world in our sample (Iran, Pakistan, and Turkey) are known to have a predominantly Islamic population, which does not preclude the presence of Islamic patients in the sample of western countries though.

Limitations

The greatest limitation of this review was the low quality of the records found for prevalence studies and the high heterogeneity. The vast majority of studies failed to adequately characterize their participant samples. We believe that the average age, gender, ethnicity, marital status, schooling, reason for rhinoplasty (esthetic, functional, or both), and percentage of patients who underwent previous rhinoplasty are essential to facilitate comparisons between subsequent studies. However, the most serious error was the lack or underestimation of the sample calculation. Our ideal estimated sample was 246 individuals, based on an expected prevalence of 20.1% (Veale et al., 2016) and 95% CI. Only one of the 21 studies selected met this criterion (Alavi et al., 2011), and none of them presented a confidence interval for the estimated prevalence. An important factor that probably complicates the homogeneity of studies is the large number of different instruments used to identify BDD. Because the lack of active participation by a surgeon tends to result in smaller samples and higher refusal rates (Crerand, 2003; Veale et al., 2003), a single easy for surgeons to deploy and interpret protocol is needed.

Conclusions

This meta-analysis found that the prevalence of body dysmorphic disorder and dysmorphic symptoms in rhinoplasty preoperative patients is around 34%, and goes down to 23% in primary rhinoplasty candidates. Prevalence rates for BDD are higher among patients who seek rhinoplasty from plastic surgeons compared to those who seek this procedure from otorhinolaryngologists. Similarly, the prevalence of BDD tends to be higher among Brazilian rhinoplasty candidates than patients from the rest of the world. Regardless of the scenario, the prevalence of BDD was markedly high, which highlights the importance of routine screening in a systematic way in rhinoplasty clinics worldwide and multidisciplinary approach. It is important to consider including psychiatrists or psychologists in the preoperative rhinoplasty team.