Search strategy

The research question guiding this integrative review was ‘What is the available scientific evidence on possible associations between sexuality-related issues and the symptoms of ED’?

In order to clarify the components that guided the search for evidence, the research was based on the PICO strategy - Participants (P), Interventions (I), Comparisons (C) and Outcomes (O) (Akobeng, 2005; ^{Cesnik & Santos, 2012}; Santos, Pimenta, & Nobre, 2007; ^{Stillwell, Fineout-Overholt, Melnyk, & Williamson, 2010}), as follows: “In individuals diagnosed with ED (P), evaluated through the application of standard scales, questionnaires and semi-structured interviews (I), how the symptoms of ED are associated to their experience of sexuality and femininity (O), in comparison with normal population (C)?”

We followed the steps recommended by the PRISMA Statement - Preferred Reporting Items for Systematic Reviews and Meta-Analyses (^{Moher et al., 2009}), namely: (1) systematic survey of international publications on the subject of sexuality and ED; (2) identification of the authors, methodological characteristics of each study, year, publishing journal, country of origin, language in which they were written, objectives and obtained results; and (3) conducting descriptive analysis of the results of the selected studies and critical evaluation of the contributions offered by the available scientific literature on the subject. Next, information regarding eligibility criteria, information sources, search strategy and study selection will be provided, according to PRISMA checklist (^{Moher et al., 2009}).

Consulted databases were: PubMed (Medical Publications), CINAHL (Nursing and Allied Health Literature), PsycINFO (American Psychological Association, literature in behavioural sciences and mental health) and LILACS (Publications from Latin America and the Caribbean). The following descriptors were selected, according to DeCS - Health Sciences Descriptors and MeSH - Medical Subject Headings, combined with the boolean operator “and”: eating disorders, femininity, sexuality and sexual behaviour. The search using the aforementioned descriptors was performed in March 2018.

The inclusion criteria for the primary studies were those addressing sexuality-related issues in the context of ED published between January 2008 and March 2018, with the following classifications: experimental studies, including randomized controlled clinical trials; cohort study; case-control study; quasi-experimental studies; non-experimental/cross-sectional studies; studies using qualitative approaches. Studies published in Portuguese, English and Spanish were selected. Therefore, the evidence levels (EL) 2, 3 and 4 were considered (^{Melnyc & Fineout-Overholt, 2005}). The established exclusion criteria were studies published in formats other than peer-reviewed articles, and articles addressing different topics from the scope of the present study. Duplicated studies were computed only once. Articles’ methodological design was defined based on the criteria proposed by ^{Sousa, Pimenta and Nobre (2007}). Searches were performed by three researchers with experience in applying the procedure.

Data abstraction and synthesis

The flowchart (Figure 1) describes the identification route, selection and inclusion of selected primary studies. For data extracting, selected articles were retrieved in full and subjected to an analytical reading. Aiming to define the eligibility of each study, selected articles were independently and carefully read by the three qualified researches (judges) previously mentioned. The level of inter-judge concordance was calculated using the formula: agreement = S possible concordance subtracted from real discordance divided by the number of articles (^{Machado et al., 2012}). To ensure the reliability of classification, the results that obtained consensus between the three evaluators were accepted automatically. The discordant items were discussed individually until evaluators came to mutual agreement.

Figure 1 Flowchart of the review process based on the PRISMA model (^{Moher et al., 2009}). 

Figure 1 shows the review process. Initially, 579 articles were retrieved. After checking for duplication, 467 were screened and 104 were excluded based on inclusion and exclusion criteria previously described. Next, 363 full-text articles were assessed for eligibility, and 344 were excluded for addressing other thematics. Finally, 19 full-text articles composed this review.

Selected articles were qualitatively analysed through thematic categorization (^{Braun & Clarke, 2006}) and will be presented in the Discussion section. Categories were defined through convergence of themes. Articles do not appear in more than one category.

1. Gender role, female sociocultural standards and ED

This category includes 11 studies that addressed gender roles and sexual behaviour in the context of ED, starting from the premise that gender role - i.e. how an individual stands in the world, within male and/or female dimensions - offers more in depth explanations to disordered eating behaviors rather than biological sex. These premises are based on those described by John Gagnon and William Simon in their sexual script theory (^{Gagnon & Simon, 1973}), which proposes human sexuality as consequence of the interaction between three levels of scripting: cultural, social and intrapsychic. In this way, sexual script theory relativizes the biological domain and broadens the understanding of human sexuality (^{Wiederman, 2005}).

Six studies overall indicated that femininity was associated to eating pathologies, and therefore configured a risk factor for the development of ED (^{Cella, Iannaccone, & Cotrufo, 2013}; ^{Mason, Lewis, & Heron, 2017}; ^{Matthews-Ewald, Zullig, & Ward, 2014}; ^{Lampis et al., 2017}; ^{Watson et al., 2017}; ^{Yean et al., 2013}). These studies started from definitions of femininity and masculinity recommended by the Bem Sex Role Inventory (BSRI), which is a widely used self-administered instrument for measuring gender role identification. It entails that socially desirable feminine characteristics are: affectionate, loyal, eager to soothe hurtful feelings, soft spoken and tactful. On the contrast, socially desirable masculine characteristics involve: independent, makes decisions easily, willing to take risks, acts as a leader and competitive. It is worth emphasizing the need for updating these definitions, due to existing social movements on behalf of gender equality (^{Butler, 2011}; ^{Wiederman, 2005}). We consider that self-defined, non-essentialist and pluralist ideas of gender identity are more in line with the current social scenario.

^{Cella, Iannaccone and Cotrufo (2013}) and ^{Lampis et al., (2017}) developed similar studies in which they aimed to investigate possible associations between gender role and eating pathologies. Both studies found that stereotipically feminine traits were linked with high levels of ED symptoms, especially bulimia, drive for thinness and body dissatisfaction. Moreover, adolescents whose gender role was socially considered discrepant from their biological sex were more likely to show higher levels of bulimia and drive for thinness. ^{Lampis et al., (2017}) discuss the influence of historical, political, economical and religious context, considering that the binary conceptualization of gender is still strong in Italy, where their study was developed, and contributes to the dissatisfaction with one’s own body.

Still considering the impact of social standards on women’s eating attitudes and body (dis)satisfaction, ^{Woolhouse et al., (2012}) developed a qualitative study that aimed to investigate reports of young women about food, eating, femininity, body and the internalization of traditionally feminine social roles. Thinness was portrayed as a symbol of social, economical and cultural success; restriction and self-denial indicated control, self-discipline, and moral virtue; fat was considered a mark of deviation, absence of self-discipline and something to be “burned” and expurgated; eating was as a source of guilt and shame, and food was sinful and an enemy to be defeated. Accordingly, ^{Green et al., (2008}) suggested that women who endorse gender roles traditionally attributed to women, such as dependency and passivity, also present an exacerbated need for social approval. Therefore, these women are at high risk for developing ED.

Results indicated that overvaluation of thinness is the most important factor in the relationship between high compliance with the traditional norms and symptoms of ED. Overall, both studies (^{Green et al., 2008}; ^{Woolhouse et al., 2012}) ratify the understanding that restrictive eating and dieting are not individual choices entirely: they are mainly outcomes to neoliberal discourses that overestimate thinness and, as a consequence, segregate women whose bodies do not fit into the norm. Thus, every woman who wishes to feel like they fit in, needs to stay “fit”.

^{Churruca, Usher and Perz (2016}) highlighted the importance of considering the broader role of sociocultural, political and gender issues in explaining ED, along with the attention that has been paid to individual risk factors. Authors aimed to examine constructions of food developed by women with BN, how these constructions relate to cultural discourses and their implications for subjectivity. Results showed six ways in which participants constructed food: “good/healthy” or “bad/unhealthy”, “contaminating body and soul”, “collapsed into fat”, “pleasurable reward”, “comfort” and “fuel for the body”. Notions of food as comfort and a forbidden pleasure provided two different ways for participants to rationalize their binging: as a form of self-care and medication, or as a disruption to their self-control. Thus, despite women deploying these constructions to make sense of their bulimic behaviours, they are culturally normative, which has implications for therapeutic and preventive strategies for BN.

^{Yean et al., (2013}) and ^{Watson et al., (2017}) found an association between male homosexuality, dysfunctional eating behaviours and body dissatisfaction. These results were explained due to the presence of feminine traits in homosexual men, as defined by the BSRI. Hence, masculinity was considered a protective factor against the occurrence of ED. ^{Zullig, Matthews-Ewald and Valois (2017}) brought similar results: individuals with high levels of femininity (including homosexual males) showed more restrictive eating behaviours than individuals with high levels of masculinity.

Accordingly, the study developed by ^{Matthews-Ewald, Zulling and Ward (2014}) showed that homosexual women - who would present masculine traits (BSRI) - presented few dieting habits, which were not considered dysfunctional eating behaviours, such as vomiting or use of laxatives. However, when compared to heterosexuals, homosexual women were at similar or increased risk for ED (^{Mason, Lewis, & Heron, 2017}), which can be explained by the existing binary formulation (^{Lampis et al., 2017}) that marginalizes sexual minorities and contributes to their body dissatisfaction. Sexual orientation discrimination was directly associated with disordered eating (^{Mason, Lewis, & Heron, 2017}).

Finally, the study developed by ^{Gordon et al., (2016}) investigated the impact of societal femininity ideals on disordered eating behaviours in transgender women. It is worth emphasizing that “transgender women” comprise “those whose gender identity - such as woman, transwoman, Male to Female (MtF), or genderqueer - differs from their assigned male sex at birth” (^{Gordon et al., 2016}, p. 2). This research complies with studies previously presented in this category, since it deals with Western society’s constructions about the “ideal female body” and its influence on identity constitution, as well as its relation to symptoms of ED. Findings showed that weight and shape control desires and behaviours were situated at the intersection of four emergent themes: (1) gender socialization and the development of femininity ideals, (2) experiences of stigma and discrimination, (3) biological processes, and (4) multi-level sources of strength and resilience. Sexual objectification, along with stigmatization, was commonly reported by participants and contributed to increase self-objectification and body consciousness. These, in turn, increased psychological strain, body dissatisfaction, and disordered methods of weight control.

The studies included in this category presented, in general, the importance of addressing sexuality, gender identity and gender role in health assistance to better understand the socio-cultural context underlying ED. Studies also associated ED to the culturally and socially imposed ideal of beauty as synonym of personal success for women. This ideal seem to rule attitudes and behaviours of individuals with feminine personality traits, but not individuals with male traits.

Four out of five studies published in 2017 approached the relationship between ED and gender role/sexual orientation, as well as the most recent studies included in this review. This shows that the social reorganization of gender role concepts that has been developed over the last decade (^{Lampis et al., 2017}) has currently been focus of attention in the scientific community, contributing to progresses in the area of sexuality, including dialogues with the ED field.

2. Attitudes towards sexuality, adverse experiences and its relations to ED

This category of analysis includes studies that overall aimed to describe sexual functioning of women with ED and the adverse experiences that may have influenced it. In the study developed by ^{Pinheiro et al., (2010}), among the 242 participants, 98% reported having experienced intimate relationships, and 55% of these relationships included sexual activities. Approximately 60% of participants with ED reported libido decrease and anxiety related to sexuality. Women with AN demonstrated higher prevalence of libido decrease than women with BN and with non-specified ED. Lack of sexual activity and libido decrease were associated with low body mass index (BMI) throughout life, high levels of interoceptive consciousness and anxiety traits. Sexual anxiety was associated with low BMI, avoidance behaviours and inefficiency. The incidence of sexual dysfunction in women with ED was higher than the normative sample.

Accordingly, ^{Dunkley, Gorzalka and Brotto (2016}) developed a study that aimed to explore altered sexuality in relation to disordered eating, emphasizing the role of sexual insecurities. Results indicated that aspects of disordered eating were associated with more cognitive distractions during sexual activity and lower sexual self-efficacy. Measures of disordered eating significantly predicted body and performance-based cognitive distractions during sexual activity, as well as global sexual self-efficacy and all efficacy subscales. Also, women with high risk for ED reported more sexual insecurities compared to women with typical levels of ED risk.

Regarding the relationship between unwanted sexual experiences and ED, ^{Castellini et al., (2013}) found that women submitted to Cognitive Behavioural Therapy for AN and BN showed significant improvement in all domains of sexuality, except those who had history of sexual abuse. These participants showed high levels of body image concern at follow-up, which indicates that one’s relationship with the body seems to be negatively affected by sexual abuse. ^{Capitaine et al., (2011}) also emphasize that negative sexual experiences may have an impact on body image, leading to negative emotions regarding femininity and sexuality, which may precipitate the dysfunctional food restriction. In addition, ^{Tolosa-Sola et al., (2017}) uphold that even when controlling for variables including sexual abuse, personality disorder diagnosis and severe depressive symptoms, body dissatisfaction remains associated with sexual impairment among women with ED. Results from these studies indicate that both body and sexual dissatisfaction are inherently related to ED symptoms, and this relation can be strengthened by unwanted sexual experiences.

Two studies assessed mechanisms that women with ED use to cope with the emotional repercussions of sexual abuse: thought suppression (^{Collins et al., 2014}) and non-suicidal self-injury (NSSI) (^{Gonçalves et al., 2016}). Thought suppression involves attempts to remove unwanted thoughts from consciousness. However, participants showed an increase of ED symptoms after suppressing thoughts related to the experience of recent sexual abuse (^{Collins et al., 2014}). Thus, suppressing thoughts from the traumatic experience of abuse resulted in anxieties that were unconsciously headed towards food and body image. Similarly, NSSI in ED patients were associated with purging behaviours, feeling fat and sexual/physical abuse. Authors concluded that ED patients with history of sexual abuse are more likely to engage in maladaptive behaviours as emotion regulation strategies (such as NSSI) when purge symptoms are absent.

Finally, considering possible associations between parental traumatic experiences and the occurrence of ED in the offspring, ^{Cerniglia et al., (2016}) investigated psychological functioning of adolescents with BED and their parents. Females showed higher tendency to internalize feelings and emotions, and higher scores on Withdrawal subscale of the Youth Self-Report (YSR). Results also showed that females made use of Somatization and Dissociation, which means that girls with BED usually dissociate from negative emotions using food to numb internal discomfort. Interestingly, mothers of girls with BED also showed higher internalizing symptoms and higher scores on the Depression scale. Regression analysis correlated Depression with higher scores of Withdrawal and Somatization, which suggested links between same-gender dyads (mother-daughter). It is also noteworthy that most of traumatic experiences in mothers resulted from physical and/sexual abuse. This shows that traumatic experiences generate negative psychological reverberations that may persist into the next generation, especially between family members who share the same gender. Several current studies have addressed the issue of psychic transmission among generations of families of members with ED (^{Leonidas & Santos, 2015a}, ²⁰¹⁷; ^{Moura, Santos, & Ribeiro, 2017}; ^{Valdanha-Ornelas & Santos, 2016}).

Limitations of this review

This review presents several limitations, mostly the lack of randomized controlled trials, which would enable the establishment of possible cause-effect relationships to better understand the association between sexuality and ED. However, sexuality is a multidimensional construct that can be discussed through a wide range of approaches, such as social, psychological, behavioural and biological. Thus, cause-effect relationships should be put in perspective, discussed and articulated along with the subjective and experiential aspects of this construct.

Finally, most of the studies included in this review were based on reductionist conceptions of femininity and masculinity, approaching typical patterns of the normative cisgender phenomena. Therefore, it is worth emphasizing the importance of considering the plurality of the subjective forms of experiencing sexuality.