Personalized character of transgender medicine as a barrier to development of evidence-based clinical practice guidelines on gender incongruence

Introduction

Gender reassignment (gender transition) is a complex social, medical, and legal issue. Many, though not all, transgender people use medical interventions to bring their bodies up to the standards and characteristics of the gender which they identify with. To standardize and improve the quality of medical services in the area of gender transition in Russia, clinical guidelines are being developed [1]. Such recommendations should be based on the principles of evidence-based medicine, i.e. the effectiveness and safety of the recommended interventions should be proven by high-quality epidemiological studies.

Failure to consider non-medical (social) factors has previously been cited as a reason for the poor quality of available research and, as a consequence, an obstacle to the development of clinical guidelines [2]. However, the problem is much broader. The international guidelines issued by the World Professional Association for Transgender Health (WPATH) emphasize the individualized nature of the medical interventions required by transgender patients [3]. While some people only need to take hormonal medications, others need multiple surgeries, and still others require neither. Thus, all transgender medicine is essentially personalized medicine. The application of standard epidemiological methods in this situation is difficult because of the strong non-heterogeneity of the sample. Although evidence-based and personalized medicine may be perceived as mutually incompatible, pharmacoepidemiology proposed a number of approaches to study the variability of response to a drug that allow for the stratification of samples [4]. However, in order to stratify a sample, it is necessary to have at least a general understanding of the mechanisms of individual response. In the field of transgender medicine, although the effectiveness of medical interventions in treating gender dysphoria has been proven [5–7], the mechanisms of such effects are not exactly known and are barely discussed in the literature. The purpose of this review is to clarify the mechanisms of the effects of hormonal medications and surgeries on gender dysphoria and gender incongruence, discuss the difficulties encountered in designing epidemiologic studies due to the personalized nature of transgender medicine, and discuss the possibilities of measuring outcomes and intermediate factors for epidemiologic study design.

Mechanisms of influence in epidemiological research on transgender medicine

The diagnosis “gender incongruence” was introduced in ICD-11 to replace the diagnosis of transsexualism (ICD-10), referring to the difference between gender identity (self-perception as female, male, or another gender) and biological sex. A related term, “gender dysphoria”, refers to “the distress that can accompany the discrepancy between a person’s perceived or expressed gender and the sex assigned to them at birth”, according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) version 5. The term “gender incongruence” is broader than “gender dysphoria”, it includes other etiologies of “incongruence”, such as “gender euphoria” [8]. Some transgender people resort to medical procedures not for psychological reasons, but to conform to accepted standards of appearance [9] or (in the case of transgender sex workers) to attract clients [10]. Also, in many countries, medical interventions are still a prerequisite for civilian sex reassignment [11], so that people who may not need them for psychological reasons resort to them. Nevertheless, in most epidemiological studies, the term “gender dysphoria” is more established and used as the dependent variable.

Medical procedures for gender transition (MPGT) aim to suppress the primary and secondary sex characteristics of the original sex and to shape the characteristics of the desired sex, which, in many cases, leads to a reduction of gender dysphoria. The main MPGT include hormone replacement therapy (HRT) and surgery.

Generally, if one considers modern pharmacology rather than traditional medicine, pharmacotherapy includes drugs with the known molecular mechanisms of action on biological targets leading to the desired response. This is equally true when treating both physical and mental health issues. At the same time, most surgical procedures are aimed at correcting diseased areas of the body and, as a result, reducing suffering. In this regard, gender incongruence is a highly atypical problem, because, first, HRT does not affect gender dysphoria on a molecular level (in the same sense that, for example, antidepressants affect depression), and second, operations are performed on biologically healthy organs, the presence of which does not cause physical suffering. The action of MPGT is mediated by the changes in primary and secondary sexual characteristics. Their perception by the person him/herself, on the one hand, and, by others, on the other hand, leads to the reduction of gender dysphoria and improvement of psychosocial outcomes [1]. It is possible to propose the following scheme of the MPT impact on gender dysphoria (Figure 1).

It is useful to break down the concept of gender dysphoria into two more specific ones. In the traditional (biomedical) model, gender dysphoria is associated with a person's aversion to his or her gender characteristics [12]. However, along with this, a significant contribution to distress is made by the reactions of others to a person's perceived gender, such as using a grammatical gender that does not correspond to the gender the person identifies with (misgendering), verbal insults, physical and sexual violence. Many transgender people point out that they do MPGT not “for themselves”, but in order to be perceived by others in the right gender [9]. Thus, bodily gender dysphoria comes from the rejection of gendered features of one’s own body (genitalia, appearance), while social gender dysphoria comes from the reaction of others to a person's gender. Epidemiologically speaking, gender traits, others' perceptions, and their reactions are mediators between exposure (MPGT) and outcome (gender dysphoria). As will be shown below, the effect of exposure is modified by socio-cultural factors.

In the epidemiological studies available to date, associations are looked for only between interventions (MPGT) and outcomes (gender dysphoria and other psycho-social outcomes, such as depression and anxiety). At the same time, the authors do not take into account the mediators indicated in the chart, which leads to a highly heterogeneous sample and, consequently, a low methodological quality of research.

The remaining part of the article is devoted to a discussion of intermediate exposure mechanisms and the possibility of measuring intermediate outcomes, the consideration of which will allow for more qualitative research in this area in the future.

Sexual dimorphism and its measurement

Cisgender (non-transgender) women and men have statistically significant differences in a large number of primary and secondary sex characteristics. In particular, sexual dimorphism is present in the skeleton: the structure of the pelvis and thorax, height, limb length, and shoulder width. There are also differences in the facial structure: women are characterized by smaller and more rounded faces, a smaller nose, and a shorter upper lip, while men are characterized by pronounced brow arches, a sloping forehead with a high hairline, and more sharply outlined jaws [13]. There are differences in muscle mass, hair growth patterns, and distribution of fatty tissue. For example, women have more subcutaneous fat, which is present in the buttocks and thighs, whereas men have more visceral fat in the abdominal region [14]. Subcutaneous fat also forms fat deposits in the face [15] and breasts [16]. There are also gender differences in voice parameters [17].

The described characteristics are, for the most part, measurable. A metric of Euclidean distances between points (e.g., the width of the face at the level of the mouth, length of the nose, and distance between the eyebrows) can be used to describe the facial structure numerically [18]. Facial and body hair density is measurable using both visual scales and instrumental methods [19]. Fat distribution in individual body parts is measured by computed tomography, magnetic resonance spectroscopy, and tomography [20]. Voice characteristics such as fundamental frequency, formant frequencies, or loudness are also measurable [17]. Thus, it is theoretically possible to construct a femininity-masculinity scale by introducing these measurement results into the equation with some coefficients (more about the coefficients below).

Mechanisms of MPGT effect on sex characteristics

Feminizing HRT usually consists of antiandrogens (cyproterone acetate, spirolactone, analogs of gonadotropin-releasing hormone, finasteride, etc.) and estrogen preparations (estradiol valerate, etc.). Use of these drugs leads to the redistribution of fat according to the female type, growth of breasts, and thinning of the skin [21]. Masculinizing HRT consists of testosterone preparations and leads to a redistribution of fat according to the male type, an increase in muscle mass, facial hair growth and scalp baldness, and coarsening of the skin [22].

Redistribution of fat is one of the noticeable results of HRT. Feminizing HRT increases subcutaneous fat, masculinizing HRT decreases subcutaneous fat, while the amount of visceral fat does not change in both types of HRT [23]. Thus, in transgender women, fat deposits appear in the buttocks, hips, chest, and face, while in transgender men, on the contrary, they decrease. Similar patterns are typical for cisgender people, who have correlations between blood testosterone concentration and the waist/hip index [24]. Post-menopausal women experience accumulation of visceral fat, with exogenous estrogen replacement preventing this trend [25]. Postmenopausal estrogen deficiency also leads to atrophy of facial subcutaneous fat tissues [26].

However, the degree of fat redistribution, and therefore the effectiveness of HRT, varies from patient to patient. Studies in the general population suggest that fat redistribution may be due to genetic factors. For example, the A → G (XbaI) polymorphism in the estrogen receptor α (ERα) gene is associated with increased total fat mass and the waist/hip ratio in middle-aged women. In contrast, the rates were lower in the group of elderly women with the GG genotype than in women with AA and AG [27]. The s2431260 and rs217589 polymorphisms were associated with reduced waist circumference [28]. A genome-wide association search revealed polymorphisms in other genes associated with the waist/hip index [29, 30]. Epigenetic mechanisms also influence adipose tissue growth [31]. The described studies are aimed at identifying factors of obesity and related health problems. However, their usefulness in relation to HRT in transgender people is low, because they do not report anything about facial and breast fat redistribution, which plays an important role in gender attribution. There were also studies of genetic determinants of mammary gland growth in experiments on mice [32], but their usefulness for transgender medicine is similarly low. In addition, ethnicity can influence the effectiveness of HRT in feminizing or masculinizing the body [33].

When the effects of HRT show to be insufficient, transgender people resort to surgery. Surgical interventions include genital surgeries (orchiectomy, vaginoplasty for transgender women; hysterectomy, metoidioplasty, phalloplasty for transgender men), mammoplasty for transgender women and mastectomy for transgender men, and facial plastic surgery. The most common facial feminization surgeries are: frontoplasty (reduction of the brow arches), mandibuloplasty (reduction of the lower jaw), rhinoplasty (reshaping the nose), and chondrogaryngoplasty (removal of the Adam’s apple) [34]. Masculinization surgeries are less common and include augmentation of the brow arches, upper and lower jaw, and creation of an artificial Adam’s apple [35]. Unlike transgender men, whose voice is decreased by testosterone, the voice of transgender women does not change when taking feminizing HRT, which is why some people seek vocal cord surgery [36]. Masculinizing HRT most often increases facial hair growth in transgender men, but if this is not enough, beard transplantation is an option [37]. Feminizing HRT can reduce body and facial hair growth, but most transgender women use waxing and other dermatologic interventions [38]. Both pharmacological (minoxidine, finasteride) and surgical interventions (growth line shift, hair transplantation) are used to combat alopecia [39].

Insufficient understanding of the biological factors influencing the effectiveness of HRT in feminization and masculinization makes it impossible to predict the required dosages of medication to achieve the desired result in each individual case. Besides. it is still impossible to say in which patients, HRT will be sufficient, and in which patients, additional surgery will be required.

Measuring gender dysphoria

A large number of experimental-psychological diagnostic tools were proposed to measure gender dysphoria [40]. At the same time, the term “gender dysphoria” attracts much criticism. What is important for further discussion is not that the presence of dysphoria is used to access gender reassignment and the corresponding diagnosis had long been classified as psychiatric [41]. There are questions about the validity of gender dysphoria scales. To determine discriminant validity, scales must be tested on two groups: people diagnosed with transsexualism/gender incongruence/gender dysphoria and a control group (e.g., [42]). However, the objectivity of the criteria for these diagnoses is questionable. Often patients tell the doctor a story that would allow them to receive a diagnosis and, as a consequence, the possibility of sex reassignment: that they were born in an “alien” body, have felt dysphoria since childhood, have played with toys designed for children of the opposite sex, and so on [43]. Diagnoses do not take into account the existence of people with non-binary gender identities (who do not identify themselves as women or men), as well as people for whom dysphoria is not the main motive for gender transition. The use of different diagnoses with different criteria in different countries (for example, “transsexualism” in Russia and “gender dysphoria” in the United States) does not allow speaking fully about the identity of the samplings used in the different studies [44]. All this does not allow speaking about the high quality of epidemiological studies that use gender dysphoria as an outcome. A way out could be the proposal of scales that take into account different (including non-pathological) motivations for gender transition.

Mechanisms of the sex feature impact on gender dysphoria

In the above scheme, sex features do not influence gender dysphoria directly. In the case of social gender dysphoria, it is mediated by the perceptions of others and their reactions to the appearance. Below, the mechanisms of attribution of gender, as well as the possibilities of measuring these intermediate outcomes, are discussed in more detail.

Attribution of gender in controlled experiments

At the beginning of the article, the differences between women and men are described from a positivist perspective as measurable and objective. However, the situation becomes more complicated when a human observer is introduced into the system. It turned out that in the task of attributing gender (categorizing a stranger as female or male), not all of the dimorphic attributes described above had the same contribution. Katcher (1955) [45] used drawings of people with mixed “gender cues” to understand their degree of importance for gender attribution by 3-9-year-old children. For example, in all other respects, a “male” figure was drawn with breasts and long hair and a “female” figure was added with male genitalia. The gender cues were arranged in the following descending order of importance: clothing, hair, genitalia, and breasts. Thompson and Bentler (1971) [46] repeated the experiment, using dolls, on a sample of children and adults. Hair and body shape were the most important attributes of gender for children, while for adults, it was genitalia and body shape. Further research revealed that “feminine” and “masculine” traits contributed unevenly to gender perceptions. For example, the presence of the penis in a drawing led to the attribution of the male gender in 96% of cases, whereas only two-thirds of respondents relied on the vulva image to attribute the female gender [47]. This difference can be explained from an evolutionary point of view by the fact that it is more dangerous to confuse a male with a female than vice versa [48].

However, since people in most cases do not walk around naked, the structure of the external genitalia is not such a significant factor in attributing gender to transgender people, except in certain situations, such as going to the pool, the gym, or the doctor. On the other hand, the face is the key element of appearance that many people do not hide (except when wearing a niqab or mask). Ascribing gender (along with other characteristics such as race, ethnicity, and age) by facial features is one area of research on face perception [49]. In one experiment, participants were asked to classify individual facial features by gender. As a result, facial features were ranked in the following descending order of importance: eyebrows and eyes, eyebrows separately, eyes separately, whole jaw, chin, nose and mouth, and mouth separately [50]. Another study showed that the upper third of the face contributed the most to gender attribution [51]. Additional elements involved in attribution may include skin texture [52] and lip color [53].

Knowing which gender features play the greatest role in attributing gender would theoretically assign to them coefficients which they would enter the feminization-masculinization scale with. However, this is complicated by the presence of social factors, discussed below.

Gender attribution in the real world

The face experiments described in the previous section were conducted under tightly controlled conditions. The images were either drawn or derived from photographs, wherein any additional elements that might signal gender, such as hair, were removed. However, real people can have both facial and scalp hair, use makeup, and wear clothes that are perceived as female, male, or gender-neutral. In everyday situations, we do not necessarily see people in full-face, as in the experiments described, and in different lighting conditions. In addition, gender attribution takes place in a socio-cultural context, so the result of this classification will depend on the meanings attributed to gender in a given society. For example, in a society where men are expected to wear beards, an androgynous man without a beard is more likely to be perceived as a woman than in a society with looser gender norms. On the other hand, in a society with higher requirements for femininity, an androgynous man is less likely to be perceived as a woman. As Kessler and McKenna [47] note, “once gender attribution has occurred, everything a person does will be viewed according to that gender attribution. For example, if a person has been recognized on the basis of other gender attributions as female, but is tall and speaks in a low voice, she will be perceived as a woman who is tall and speaks in a low voice, rather than as a man”. That said, the knowledge that there are people who do not fit into the gender binary is an additional factor that influences the attribution of gender in situations where gender traits are not aligned with each other. Transfeminist Julia Serano noted that at a time when she still identified as a male crossdresser, she was more often perceived as a woman in the suburbs, whereas in larger cities, whose residents were more aware of people like her, she was more often seen as a crossdresser [54].

Reaction of others and individual response

So, based on the sum of the gender features, a person can be classified by others as unambiguously female, unambiguously male, or a person of an unclear gender. The impact on gender dysphoria and psycho-sexual outcomes for this individual now depends on how those around them behave in this situation. If a person is read as crossdresser, transgender or androgynous, negative reactions are possible, ranging from disapproving glances to verbal abuse and physical violence [55]. In other cases, a person may be perceived as a representative of one of the binary gender categories, but identify with another (if no transition has been initiated), which leads to misgenderism, i.e. the use of grammatical gender, pronouns, or address in relation to a person that does not correspond to their gender identity [56]. Such reactions naturally lead to an increase in social gender dysphoria, levels of depression and anxiety, and low self-esteem. On the contrary, if a person is read as a person of the gender he or she identifies with, this will contribute to higher self-esteem and self-confidence. At the same time, the same reactions of others can lead to individual mental health consequences depending on stress tolerance, which for transgender people varies widely and depends, among other things, on such social factors as the availability of social support [57].

Although the reactions of others and the response to them are individual in each case, opinion polls can be used to determine in which countries a transgender person has a higher risk of being attacked. For example, a 2016 survey in 23 countries showed that the highest level of acceptance of transgender people was in Spain, while the lowest was in Russia [58]. Thus, social attitudes are an additional variable influencing the reactions of those around them, which, in turn, increases or decreases social gender dysphoria.

Conclusions

Researchers wishing to establish associations between medical procedures of gender transition and gender dysphoria face many obstacles due to a lack of understanding of the mechanisms of impact and the difficulty (sometimes impossibility) of measuring intermediate outcomes. The experience of gender dysphoria is individualized, and the validity of the scales used to measure it is questionable. People’s appearances prior to gender transition are different, so some need more intervention to be perceived in their gender than others. The effects of exogenous hormones on the appearance are individualized and possibly influenced by genetic factors, but studies for transgender people are lacking. Attributions of the same person’s gender can vary in specific situations and depend on the social context. The reactions of others to a person’s perceived gender are individual and depend on the social context. Additional variables can be introduced into regression equations to overcome these complexities. For example, a feminization-masculinization scale could theoretically be developed based on anthropometric measurements and coefficients taken from studies showing the different significance of gender features in gender attribution. Pharmacogenetic studies could be conducted to determine the genes that affect the degree of feminization or masculinization with hormone therapy. Additional research could be conducted on the contribution of the socio-cultural context to gender attribution and the reactions of others. Transgender medicine is actively developing all over the world, and one can expect to see studies on these topics in the coming years.