The Relationship Between Anabolic Androgenic Steroid Use and Body Image, Eating Behavior, and Physical Activity by Gender: A Systematic Review

The prevalence of anabolic androgenic steroids (AAS) is rising, especially in recreational sports and the general population. While body image significantly influences AAS use, gender differences remain unclear. We examined gender-related connections between AAS use, body image, eating behavior, and physical activity. Following PRISMA guidelines, we analyzed 22 studies: 14 with male-only samples, 5 mixed-gender, 2 with sexual and gender minorities, and 1 with a female-only sample. Findings confirm body image as a key predictor of AAS use. Though AAS use correlates with eating disorders, outcomes vary by context; for instance, no discernible difference in eating behavior was observed between AAS users and non-users in bodybuilding. Physical activity findings varied, with some studies showing no significant differences between AAS users and non-users. Due to limited gender-comparison studies, conclusive gender-related differences cannot be drawn.


Introduction
Anabolic androgenic steroids (AAS) are a class of compounds containing chemical as well as (semi-)synthetic derivatives of testosterone, the primary male sex hormone (Fadah et al., 2023;Pope & Brower, 2009).Aside from potential therapeutic benefits (see e.g., Basaria et al., 2001;Brock et al., 2016;Rodrigues dos Santos et al., 2020), AAS have been predominantly utilized within professional sports contexts, e.g., by competitive weightlifters and bodybuilders seeking to augment muscle mass, definition, and strength to improve athletic performance (Ip et al., 2011), despite their use being prohibited (World Anti-Doping Agency, 2024).AAS use in such cases is mostly non-medically prescribed and often exceeds physiological levels or clinical recommendations (Kanayama et al., 2008).
Since the 1970s, there has been a notable shift in AAS use patterns, with prevalence expanding beyond organized professional sports (Bates et al., 2019;McVeigh & Begley, 2017).There is evidence that AAS are increasingly being used by the general population, recreational athletes, and regular gym-goers to improve both physical appearance and performance (Ip et al., 2011;Kanayama & Pope, 2018).According to data from the US national household survey from 1991, the general population's lifetime prevalence estimates for the use of AAS were 0.9% among men and 0.1% among women (Yesalis, 1993).A comprehensive meta-analysis in 2014, however, estimated a lifetime prevalence of 3.3% in the general population, with a significantly higher prevalence observed among men (6.4%) compared to women (1.6%; Sagoe et al., 2014).Furthermore, estimated prevalence soar within specific subpopulations, such as individuals with substance use disorders (8.0%) and athletes J o u r n a l P r e -p r o o f (13.4%), reaching as high as 18.4% among regular gym attendees (Sagoe et al., 2014).However, it is noteworthy that the actual prevalence of AAS use may likely be substantially greater, as these estimates are typically derived from self-report questionnaires or interviews, inherently susceptible to social desirability bias stemming from stigma, especially in women (Havnes et al., 2021).
The misuse of AAS can lead to a range of severe and detrimental physical, psychological, and social consequences, including the development of cardiomyopathy (Fadah et al., 2023), a negative impact on male fertility (Mulawkar et al., 2023), and possible brain atrophy (Bjørnebekk et al., 2017).Furthermore, studies show a connection between AAS use and aggression/violence (Hauger et al., 2021;van de Ven et al., 2023).Additionally, AAS use has been observed within the context of substance abuse and suicidal behavior (Ganson & Cadet, 2019;Lindqvist Bagge et al., 2017;Pope et al., 2014).For sexual minorities, Ip et al. (2019) found that gay and bisexual men who use AAS are more prone to unsafe sexual behaviors than their non-using counterparts.The same authors reported an increased risk of infectious disease transmission among AAS users in this population (Ip et al., 2019).
Additionally, Nagata et al. (2022) found an association between the number of sexual partners and AAS use in cisgender gay men.
Body dissatisfaction and body dysmorphia may play a central role in motivating individuals to use AAS (e.g., Bonnecaze et al., 2020;Goldman et al., 2019;Olivardia et al., 2000;Pope et al., 2017).Recently, potential causal factors for the development of AAS use have been identified by applying machine learning, including symptoms of muscle dysmorphia and body dissatisfaction (Hudson et al., 2023).Griffiths et al. (2018) suggest that increasing numbers of men are concerned about their body and muscularity in response to exposure to fitness-related media.In a similar manner, the societal portrayal of the ideal female physique, which now accentuates muscularity, has also changed noticeably in recent decades (Fasoli & Constantinou, 2024).Despite this emergent trend, empirical data on AAS use among women remain relatively scarce (Scarth et al., 2022).
A well-documented nexus exists between body image and both physical activity and disordered eating behavior (Strand et al., 2021).Body image disturbances are in fact core symptoms of eating disorders (ED) (American Psychiatric Association, 2013).Adolescents with a positive body image tend to exhibit healthier levels of physical activity, healthier eating habits, and higher self-esteem, while a negative body image may go along with excessive or compulsive exercise and disordered eating behavior (Kopcakova et al., 2014;Pope & Brower, J o u r n a l P r e -p r o o f 2009).Although ED have historically been predominantly associated with women (Halbeisen et al., 2024), newer research suggests that up to one in four clinical cases of ED could involve men (Galmiche et al., 2019;Santomauro et al., 2021).The drive for muscularity is associated to ED pathology, especially in men (Lavender et al., 2017), but also in women (Cunningham et al., 2019), thereby also establishing a link between the use of AAS and ED (Goldfield, 2009;Goldfield & Woodside, 2009).
In summary, the prevalence of AAS use appears to be on the rise, accompanied by a spectrum of well-documented adverse effects.The surge in AAS use underscores the imperative for a comprehensive understanding of this phenomenon.Such insight is paramount for devising effective interventions and mitigating the onset of enduring and severe health sequelae.Furthermore, there is a need for additional exploration into the intricate relationship between AAS use and gender.Body image and its behavioral correlates of eating behavior and physical activity emerge as pivotal factors in this realm.This systematic review aims at exploring the relationships between AAS use and aspects of body image, eating behavior, and physical activity and at providing insights on gender differences within this context.

Search strategy and selection of primary studies
The review was pre-registered with PROSPERO (CRD42023453390) and was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidance (Moher, 2009) (see Figure 1).An evidence-based electronic search was conducted in accordance with the Peer Review of Electronic Search Strategies (PRESS) guideline for systematic reviews using four databases (Scopus, Web of Science [WOS], PubMed MEDLINE and PsycINFO; last search August 15, 2023).For each of these databases, a complex search strategy was developed consisting of a combination of MeSH terms, keywords, and text words in different spellings connected to AAS use, gender, eating behavior, body image, and physical activity (for Scopus, see Table 1; search strings for all databases see https://www.crd.york.ac.uk/PROSPEROFILES/453390_STRATEGY_20230810.pdf).
The publication period was limited to the last ten years (2013 onwards) to ensure that prevalence estimates are the most up to date.No limitations were imposed regarding the country of origin of the sample or ethnic groups, and all types of AAS use were considered without restriction.After importing the search results into the Covidence systematic review software (Veritas Health Innovation, 2024), all duplicates were removed, and a title/abstract-J o u r n a l P r e -p r o o f screening was carried out by seven (CZ, NML, RM, KA, LS, CB, GB) independent reviewers, with each study rated independently by two reviewers.Year, language AND PUBYEAR > 2012 AND PUBYEAR < 2025 AND (LIMIT-TO (LANGUAGE, "English")) Note.AAS=anabolic androgenic steroids.

Inclusion and exclusion criteria
Inclusion and exclusion criteria are listed in Table 2 according to the relevant domains.In case of disagreement at the initial stage of title and abstract screening between the ratings of the two reviewers per study, conflicts were discussed with all reviewers and a consensus decision was reached for inclusion or exclusion.After that, all full texts were retrieved, and each was screened and rated by two independent reviewers.Conflicts regarding the rating of the full texts were discussed among at least three independent reviewers, and consensus was reached for the inclusion or exclusion of the study under the supervision of the first author (CZ).The final selection was approved by the entire study team (i.e., all authors).

Table 2
Inclusion and exclusion criteria.

Data extraction and assessment of methodological quality
All data (e.g., demographics including country of origin of the study sample, study design, time of data collection, and outcomes of interest) were evaluated and extracted by two independent reviewers (for details see Supplementary Table 1).In case of discordance, the first author (CZ) was consulted, and consensus was reached by discussion.The risk of bias was analyzed using the "Critical Appraisal Checklist for Analytical Cross-Sectional Studies" (The Joanna Briggs Institute, 2017) as suggested by Ma et al. (2020).

Results
The systematic search revealed 9,839 studies, of which 6,510 records were retained after the deletion of duplicates (N=3,329).Following the screening of title and abstract, 248 publications remained to complete the review (see PRISMA diagram, Figure 1).J o u r n a l P r e -p r o o f

Basic characteristics and quality of included studies
Finally, a total of 22 studies were included, encompassing a total participant pool of 29,616 individuals, of whom 2,268 (7.7%) identified as individuals with either current or lifetime AAS use.Among these participants, 16,958 (57.3%) identified as men, 12,203 (41.2%) identified as women, 335 (1.1%) identified as gender-fluid, gender-queer, gender-neutral, nonbinary, transgender or "other", while 120 participants (0.4%) did not report their gender.The present review included a wide range of sample compositions, although there was a noticeable imparity towards the availability of men-only studies.Fourteen studies contained solely men, two included sexual minority populations (Griffiths et al., 2017;Klimek-Johnson et al., 2023), five studies had mixed-gender samples (Ganson et al., 2022;Gestsdottir et al., 2021;Gonçalves et al., 2023;Gwizdek et al., 2018;Nagata et al., 2022) and one study exclusively centered on women (Scarth et al., 2022).Three studies assessed only individuals using AAS (De Zeeuw et al., 2023;Murray et al., 2016;Zoob Carter et al., 2021), one study compared individuals using AAS and a control group (Scarth et al., 2023), while the proportion of individuals using AAS varied significantly in the remaining studies, ranging from 0.3% to 63.5%.The age range of participants spanned from 14 to 78 years.
Most studies were conducted in Western industrialized countries.Out of the 22 studies, 14 did not report race or ethnicity.Seven studies were from the United States of America (USA), four from Norway, two from Italy, and others from Australia and New Zealand, the Netherlands, the UK, Iceland, Brazil, Poland, and Iran with one study each, as well as one international sample (including 42 countries, primarily from the USA, UK, and Canada).
Among the studies that reported race/ethnicity data, at least 60% of participants were White/non-Hispanic or Caucasian.Diverse samples were noted only in the cohorts of Klimek-Johnson et al. (2023) and Nagata et al. (2022), showcasing broader ethnic representation.
Detailed information on the country of origin for each study is provided in Supplementary Table 1.
Furthermore, only the study by Gonçalves et al. ( 2023) included a power analysis, but the sample size did not meet the required criteria.An overview of the quality assessment is displayed in Table 3.

Table 3
Quality rating of included studies.
J o u r n a l P r e -p r o o f Note. + = yes; − = no; ?= not determined/not applicable; NR = not reported.
The results of this systematic review are organized according to the three primary outcomes (body image, eating behavior, physical activity).Statistical test results or mean values (M) and standard deviations (SD) of the questionnaires used were reported, where possible.Gender comparisons were also reported, where available.Several studies examined multiple outcomes that were considered relevant to this synthesis and were therefore enumerated for each specific outcome, resulting in dup-and triplicate entries in the results section (De Zeeuw et al., 2023;Gestsdottir et al., 2021;Griffiths et al., 2017;Gwizdek et al., 2018;Ip et al., 2015;Jenssen & Johannessen, 2015;Klimek-Johnson et al., 2023;Murray et al., 2016;Nelson et al., 2022;Scarth et al., 2023).The results are sorted by sample context: Education context (high schools, universities), exclusive samples with AAS use regardless of the context, sexual minority individuals, and sports context (athletes, bodybuilders, weightlifters).

Body Image
This review included 14 studies with a total sample size of N=19,039 (n=12,707 men, n=6,098 women, n=116 "other" gender), which could be associated with the outcome of body image.
In total, the studies examined N=1,308 current or former users of AAS.The estimated lifetime prevalence of AAS use ranged from 1.4% to 1.9% among samples from educational settings.
Additionally, one study reported a prevalence of 4.0% for current AAS use in educational contexts.Among athlete samples, the lifetime prevalence ranged from 6.7% to 43.8%, with J o u r n a l P r e -p r o o f one study indicating that 26.7% of athletes were current users of AAS.The results of the included studies regarding body image are shown in Table 4.

Table 4
Results regarding body image.

Research question
Outcome Measure

Main result regarding outcome Sample
Educational context  -Those using AAS for appearance purposes exhibited greater drive for size (p<.001, -Spearman's correlation coefficients between lifetime duration of AAS use and MDDI score were .20,for the subscale "drive for size" .10,for the subscale "appearance intolerance" .20 (only the latter being significant at p<.01).-Professional women athletes were more likely to report "being proud of their muscles" (p<.001).-Among amateur athletes, no differences in body confidence were found between men using and not using AAS.-Body confidence, along with male gender, was identified as a factor contributing to AAS use.J o u r n a l P r e -p r o o f physical problems -16.2% of AAS using individuals (vs.24.0% in the non-using group) were attempting to maintain weight.

Eating Behavior
This review included eight studies that focused on eating behavior outcomes.The studies comprised a total of N=23,398 participants (n=11,657 women, n=11,165 men, n=114 genderfluid, gender-queer, gender-neutral, non-binary, transgender, or "other" gender).
Regarding the number of participants with AAS use in the samples, two studies in educational contexts reported an estimated lifetime prevalence of 0.3% and 1.6% for AAS use, respectively.Additionally, one study reported an estimated prevalence of AAS use of 5.6% in the past year.Moreover, these studies indicated that men exhibited a higher estimated prevalence of AAS use compared to women (Ganson et al., 2022;Gestsdottir et al., 2021;Nagata et al., 2022).The results for eating behavior are shown in Table 5.
While Nagata et al. (2022) provided an estimated prevalence of AAS use within the educational population, their subsequent analyses combined AAS use with other musclebuilding substances such as creatine and amino acids.This aggregation may have obscured the specific impact of AAS use on the reported results.The sample in the study conducted by Klimek-Johnson et al. (2023) was complex in term of different groups of sexual orientation.
Nonetheless, the study did not provide any specific results comparing these groups regarding AAS use or eating behavior.

Table 5
Results regarding eating behaviors.

Research question
Outcome Measure

Physical Activity
This section includes the findings concerning AAS use and physical activity obtained from the inclusion of 12 studies.The total sample comprised 13,200 individuals (n=5,301 women, n=7,899 men) with 1,322 (10.0%) individuals currently or formerly using AAS.In the studies conducted within an educational setting, one study documented an estimated lifetime prevalence of AAS use at 1.6%, while another study reported a point prevalence of 4.0% among male students.In the studies conducted within a sports context, two studies indicated a lifetime prevalence of AAS use at 6.7% and 29.1%, with current use prevalence being 24.6% and 26.7%, respectively.The main results are summarized in Table 6.
Overall, studies have shown that the majority of people who use AAS exercise at least four times a week (Dunn & Piatkowski, 2021;Ip et al., 2015;Khadem Amiri & Salehidoost, 2023;Nelson et al., 2022).The studies by De Zeeuw et al. ( 2023) and Zoob Carter et al. (2021) reported relevant outcome results in relation to the COVID-19 pandemic.

Table 6
Results regarding physical activity.Prevalence of AAS use, mental health, and substance use Frequency of (non-) organized sports -Significant differences between individuals using and not using AAS with regard to engagement in nonorganized sports (p=.001).-Significant associations between AAS use with male gender and participation in non-organized sports.At T1, 91.1% reported some level of impact on their current training with 48.5% described it as "high" to "extremely high".At T2, 66.7% reported "some" impact of the pandemic on their training, with 13.7% indicating a "high" to "extremely high" impact.-There was a significant reduction (p<.001) in average training frequency at T1 (M=3.9;SD=1.2) compared to prepandemic levels (M=4.4;SD=0.7), and a decrease in training frequency at T2 (M=4.1;SD=1.1) compared to prepandemic levels, although the difference between T1 and T2 training frequencies was not statistically significant (p>.05).-There was a significant reduction (p<.001) in average AAS doses at T1 (M=2.76;SD=1.14) compared to prepandemic levels (M=3.31;SD=0.95);AAS doses at T2 (M=3.03;SD=1.44) were significantly higher (p<0.05)than at T1, but still significantly lower (p<0.05)than the average pre-pandemic doses.

Main findings
The presented review included 22 studies that examined associations between AAS use and body image, eating behavior, and physical activity.The estimated lifetime prevalence of AAS use in the included studies ranged from 0.3% (Ganson & Cadet, 2019) to 29.1% (Khadem Amiri & Salehidoost, 2023), with the lowest estimated lifetime prevalence (0.3%-1.9%) found in cohorts from educational contexts, and the highest prevalence reported in sports contexts (6.7% -43.8%).This is in line with previous findings (Amaral et al., 2022).
Three studies reported gender differences regarding body image and AAS use (Gonçalves et al., 2023;Gwizdek et al., 2018;Klimek-Johnson et al., 2023).Regarding the prevalence, men were found to have higher prevalence of AAS use than women.One possible rationale for observed disparities in AAS use between genders may stem from societal perceptions regarding the effects of AAS.While there is a societal inclination to admire fitness in women, an excessive development of muscle mass is often devalued (Musolino et al., 2022).

J o u r n a l P r e -p r o o f
Additionally, secondary effects associated with AAS utilization, such as voice deepening (Barkat-Defradas et al., 2021) or clitoral enlargement (Nieschlag & Vorona, 2015), are viewed negatively in women (Havnes et al., 2021).Consequently, the evaluation of the costs and benefits of AAS use in women tends to lie more on the side of costs.At the same time, it is imperative to acknowledge that the majority of studies predominantly sampled men, resulting in an unbalanced overall gender distribution.This warrants caution when interpreting data on the prevalence of use by women.
The prevalence of AAS use in the past year in sexual minority groups was reported to be as high as 35.7% among men and 25.5% among women (Klimek-Johnson et al., 2023).In a student sample, non-heterosexual men reported higher AAS lifetime use compared to heterosexual men (Ganson et al., 2022).Only four studies (Ganson et al., 2022;Gestsdottir et al., 2021;Griffiths et al., 2017;Klimek-Johnson et al., 2023) examined differences with regard to sexual orientation and AAS use.While two studies assessed gender beyond the man vs.
woman binary (Gestsdottir et al., 2021;Griffiths et al., 2017), their outcomes were not reported differentially for gender minorities; thus, caution is required when estimating in this context.Given the dearth of insight into AAS use among women and sexual as well as gender minority individuals, further research in this realm is necessary.
Societal norms that prescribe the ideal physique to be muscular and lean may drive individuals who are dissatisfied with their bodies to consume AAS to achieve these norms (Frederick et al., 2022).In this context, cultural values regarding masculinity and femininity, as well as societal perceptions of beauty standards, may intersect with country of origin and ethnicity to influence individuals' motivations for AAS use.Several studies have found increased rates of body image dissatisfaction (Jenssen & Johannessen, 2015), muscle dysmorphia symptoms, particularly in men (Murray et al., 2016;Scarth et al., 2023), and drive for muscularity (Klimek-Johnson et al., 2023) in individuals using AAS compared to nonusers.Klimek-Johnson et al. (2023) found that high motives of thinness and muscularity were significantly associated with higher symptom levels of body dysmorphic disorder, independent of gender.Body image dissatisfaction was associated with an elevated risk of AAS use (Gestsdottir et al., 2021), consistent with findings from previous studies (Olivardia et al., 2004;Pope et al., 2012).Murray et al. (2016) reported that men using AAS primarily for appearance purposes exhibited higher muscle dysmorphia-related psychopathology compared to those using it for performance reasons.
J o u r n a l P r e -p r o o f While the present synthesis shows a possible bidirectional relationship between AAS use and dissatisfaction with body image, not all studies supported this relationship as they did not show differentially elevated rates of body image-related psychopathology in AAS users.
For example, when comparing low-and high-risk groups for muscle dysmorphia, Fabris et al. (2018) found no significant difference in AAS use.Similarly, Longobardi et al. (2017) found no association between AAS intake and muscle dysmorphia, and Ganson and Cadet (2019) reported that body perception was not a significant predictor of AAS use.Furthermore, no significant differences were found in the use of AAS based on dissatisfaction with body weight neither in men nor in women in the study by Gonçalves et al. (2023), similar to results of a comprehensive systematic study by Brennan et al. (2017).A possible explanation for such findings of no associations between body image perception and AAS use could lie in the choice of populations and/or outcomes.On the one hand, in samples from the sports context, no difference may be identifiable between those with and without AAS use, if one would assume that athletes may already be dissatisfied with their body image to a great extent.On the other hand, dissatisfaction with body weight, as assessed in some studies (e.g., Ganson and Cadet 2019), may be an aspect of body image not decisive for AAS use.Thus, the pattern but also the severity of body image dissatisfaction may significantly influence the motivation for AAS use.Furthermore, it is unclear whether AAS use exacerbates pre-existing body image concerns or if pre-existing concerns prompt individuals to resort to AAS use -a further aspect to be investigated in future studies.
Examining differences in body image concerns across diverse groups may provide insight into how societal norms intersect with AAS use behaviors.For example, LGBTQ+ community members who face distinct body image pressures (Richburg & Stewart, 2024) influenced by both societal ideals and sub-cultural norms, may exhibit differential patterns of AAS use compared to heterosexual individuals (Blashill et al., 2017).Evidence further shows that cultural contexts also impact AAS use, e.g., due to cultural differences with regard to attitudes towards male muscularity (Kanayama et al., 2010;Kanayama & Pope, 2011;Thiblin & Petersson, 2005).Similarly, cultural influences may affect eating behaviors (Higgs, 2015).
Culturally mediated effects should be further explored in future.Nonetheless, it is crucial to consider the role of culture/ethnicity in the relationship between AAS use and body image, eating behavior, and physical activity.Cultural norms may serve as significant factors influencing these variables, given the ideals surrounding body image and physical appearance within different cultural/ethnic groups.Furthermore, socioeconomic status and disparities in J o u r n a l P r e -p r o o f the access to healthcare and other resources may contribute to differences in body image concerns and health behaviors in diverse populations.
Our synthesis with regards to eating behavior revealed an increased rate of ED risk and related psychopathology in individuals using AAS (Ganson et al., 2022;Griffiths et al., 2017;Klimek-Johnson et al., 2023).Only one study dealt with gender-related differences in AAS use and ED risk and showed no significant differences between men and women (Ganson et al., 2022).Due to the lack of a broader evidence base, these results cannot be generalized.
AAS users whose motivation was to improve their appearance scored higher on the subscale "weight concern" of the EDE-Q (here in a modified version; Murray et al., 2016).The EDE-Q (Fairburn & Beglin, 2008) was additionally applied in two further studies (Griffiths et al., 2017;Klimek-Johnson et al., 2023).However, concerns about the validity and reliability of the EDE-Q in men have recently surfaced, in terms that the EDE-Q may not be adequate to assess muscularity-oriented body dissatisfaction in men, as highlighted by Laskowski et al. (2023a and2023b).
The studies identified and included in this systematic review revealed no differences in eating behavior in bodybuilders with and without AAS use.This may be attributable to the inherently strict dietary regimens commonly followed by bodybuilders (Helms et al., 2014;King et al., 2021).At the same time, individuals using AAS were found to be more likely to adhere to dysfunctional dieting behaviors (Scarth et al., 2023).Several variables may be responsible for distinct findings within specific contexts.Contextual and social factors, such as social support within the bodybuilding community (Chaba et al., 2019;Hurst et al., 2000), and/or external pressure from coaches/peers (Sansone & Sawyer, 2005), and/or risk-neglecting attitudes towards drug use (Bjørnestad et al., 2014;Chaba et al., 2019)  Regarding physical activity and its association with AAS use, the focus in the included studies was on training frequency and the type of sports performed.These studies provided detailed information about the amount of time spent on training per week, in days, hours, or even minutes, which helped distinguish between those using AAS and those who were not.
Several studies aimed to identify physical activity patterns (Gestsdottir et al., 2021;Gwizdek et al., 2018;Ip et al., 2015;Scarth et al., 2023).While certain studies showed a specific focus on training routines, frequency, and volume among individuals using AAS compared to nonusers (Ip et al., 2015;Khadem Amiri & Salehidoost, 2023), others showed no significant difference between the respective groups (Nelson et al., 2022;Scarth et al., 2022), or even reported a lower frequency as well less training hours and training days in individuals using AAS (De Zeeuw et al., 2023;Jørstad et al., 2023).One study found distinct performance patterns, with AAS use being associated with higher engagement in non-organized sports (Gestsdottir et al., 2021).Resistance training aiming to primarily alter size or body shape, performed in a gym or weight training environment, appears to be the type of activity most often performed by individuals using AAS.A trend towards women using AAS experiencing more symptoms of exercise dependence could be observed in one study (Scarth et al., 2023), with evidence from other studies suggesting that exercise dependence serves as a mediator in the relationship between exercise and eating behavior-related psychopathology (Cook & Hausenblas, 2008), especially in women (Meulemans et al., 2014).It is possible that shared underlying psychological factors contribute to both AAS use and exercise dependence, as suggested by Scarth and colleagues (Scarth et al., 2023), but the evidence is limited.Further research should investigate physical activity in more detail, focusing e.g., on the level of athletes´ professionalism in sports, individual sport biographies, and individual psychological factors to detect the contributing vs. protecting factors for AAS use vs. abstinence.
Furthermore, so far, no sexual minority as well as gender minority cohort was investigated in studies on physical activity and AAS use.The observed outcomes regarding physical activity and AAS use may be summarized in two patterns: 1) Individuals with an intense, mainly muscle building-oriented, resistance training regimen seem to be more prone to the use of AAS (presumably in an attempt to overcome physical limits of performance or appearance), 2) Individuals using AAS may exhibit reduced intensity in their training regime, potentially attributed to their awareness of the supportive effects conferred by AAS.
J o u r n a l P r e -p r o o f

Limitations
This systematic review has limitations that need to be considered.The limitations arise from the large heterogeneity of the included samples in terms of participants' characteristics, assessment tools, and outcomes.As an example, the study-specific cohorts included in this review differ in most aspects such as gender representation or level of sports professionalism (amateurs vs. professional athletes).Such heterogeneity impedes the formulation of overarching conclusions and generalizations; we cannot rule out that these aspects may have influenced the outcomes to a great deal.
In terms of the assessment of AAS use, a notable limitation across all studies but the one by Jørstad et al. (2023) was the absence of a standardized approach, which led to variability in measurements and their reporting.Moreover, a critical issue arose from the lack of differentiation between lifetime and current AAS use in many studies.Consequently, the current prevalence of AAS use might be overestimated, while the association between AAS use and the investigated outcomes might be underestimated, e.g., if participants had stopped using AAS long time ago but were still categorized as current users.
Similarly, the construct of body image, eating behavior, and physical activity represent multifaceted phenomena that are operationalized differently across studies, thus warranting caution when comparing results.Variability in conceptual frameworks and assessment methods underscores the complexity of these constructs and emphasizes the need for nuanced interpretation within the context of each study's design and objectives.Additionally, the included studies predominantly adopted a cross-sectional design, which inherently limits the ability to establish temporal relationships between variables and inferences of causality.For this reason, longitudinal studies would be beneficial to elucidate the temporal relationships and causal pathways between AAS use and its associated outcomes.Moreover, future research should prioritize representative sampling designs to better elucidate patterns across different racial/ethnic/cultural and gender groups within countries.Most of the studies included in our review used convenience sampling, underscoring the need for more rigorous methods.
Additionally, it is important to acknowledge potential biases in the existing literature due to the predominance of samples from Western industrialized countries.The overrepresentation of certain groups in these samples, such as White/non-Hispanic or Caucasian individuals, may limit the generalizability of findings to more diverse populations.
Future research should strive to include more diverse samples representing various cultural/ethnic backgrounds to better understand the inferences between cultural influences, J o u r n a l P r e -p r o o f AAS use, and associated health outcomes.This will enable a more comprehensive understanding of how cultural and ethnic factors may contribute to the complex relationship between AAS use and body image, eating behavior, and physical activity across diverse populations.
Furthermore, the inability to discern clear gender differences in AAS use and its correlates is another noteworthy limitation of the reviewed studies.This limitation arises from the prevalent trend toward men-only samples in the AAS realm and the lack of emphasis on exploring gender differences in studies with mixed-gender samples.The application of findings to women remains uncertain, highlighting the need for future research to address this gap in the literature.
Despite these limitations, this systematic review provides a comprehensive overview of our current understanding of the relationship between AAS use and body image, eating behavior, and physical activity.By acknowledging these limitations, the review sets a foundation for future research endeavors to refine methodologies and expand knowledge in this area.

Conclusions
In summary, this systematic review represents an initial attempt to comprehensively synthesize gender-specific disparities in body image, eating behavior, and physical activity among individuals employing AAS in different contexts.
A notable majority of 14 out of 22 studies featured men-only samples, with one single study focusing solely on women.Studies analyzing mixed-gender samples did not adequately explore gender differences, resulting in an incomplete portrayal of gender-specific nuances.
Nonetheless, our review reaffirmed and strengthened the existing evidence implicating body image as a robust predictor of AAS use.In addition, the influence of different aspects of body image on AAS use was focused upon.The present work provides further illumination on the complex interplay between eating behavior, physical activity, and AAS use.
The estimated prevalence in the educational sector showed that young groups are also affected and that it is important to further investigate the motivation for AAS use and its consequences in younger individuals, e.g., on body image, diet, and exercise, to develop appropriate prevention and treatment programs and raise societal awareness.Our findings underscore the urgent need for future investigations that prioritize gender comparisons including all genders, sexual minority populations, and cultural and social influences, as well J o u r n a l P r e -p r o o f -Participants currently using or having used AAS (at any time in their life) -All ages, all genders -Participants with AAS use as part of therapeutic treatments J o u r n a l P r e -p r o o f
J o u r n a l P r e -p r o o f image psychopathology psychopathology (p<.001, ⴄp 2 =0.1) compared to those using AAS for performance purposes (p=.030, ⴄp 2 =0.1).
may have differential influences upon eating behavior and AAS ingestion.Individual psychological factors like perfectionism, body image concerns, and self-esteem(Bjørnestad et al., 2014;Sansone & Sawyer, 2005;Smith & Hale, 2004) may be key underlying factors in the connection between AAS use and ED, in a nuanced interplay still to be quantified.The longitudinal case-control study byNagata et al. (2022)  provided evidence that muscle-oriented eating behavior (i.e., consumption of protein drinks) in early adolescence potentially increases the risk of AAS consumption later in life.However, confounding occurred in two aspects: 1) The aggregation of AAS use with other muscle-building substances like creatine and amino acids, thus complicating the interpretation of the results specifically pertaining to AAS use and potentially obscuring the distinct impact of AAS, and 2) the classification "muscle-building behaviors" J o u r n a l P r e -p r o o f included both dietary patterns and physical activity, so it is difficult to derive clear and unambiguous conclusions from these results.

Table 1
Search Strategy.