Muscle building supplement use in Australian adolescent boys: Cross-sectional relationships with body image, weight lifting, and sports engagement


 Background: The extent and implications of muscle building protein supplement use among adolescents is relatively unknown. We aim to present the prevalence of protein powder, creatine, and anabolic steroid use in a sample of 14-16 year-old boys in Australia, and the predictors of use, and intentions to use protein powder.
Methods: Data were obtained from questionnaires with N=237 Australian adolescent boys aged 14-16 years from one independent boy’s school in Melbourne Australia. Hierarchical linear and logistic regressions were used to determine the predictors of intentions to use, and actual use of protein powder, respectively.
Results: 49.8% of boys reported current use of, and 62% intended to use protein powder; 8.4% used creatine, and 4.2% used anabolic steroids. Higher levels of participation in drive for muscularity, weight training, and playing a greater number of sports were significant predictors of higher current use and higher intentions to use protein powder, but age, BMI, body esteem, and ethnicity were not.
Conclusions: Prevalence of muscle building supplement use was relatively high among this adolescent population. This research has implications for intervention and prevention programs to educate young boys about muscle building supplements to reduce negative physical and psychological health effects of their use.

Steroids [AAS] among adolescents is relatively unknown (5,6). Research has also indicated that many muscle building products are spiked with testosterone or amphetamine-like substances, even though these are not declared on the label (7)(8)(9). PIEDs use is also associated with multiple psychological harms, including body dysmorphia and depression (10). There is also potential for a "gateway effect", whereby the consumption of relatively benign substances can escalate and lead to more serious, and illegal supplement and substance use (11,12). In support of this effect, research has found that competitive athletes were 3.5 times more likely to engage in doping if they had previously used nutritional supplements Prevalence of the use of muscle building supplements, including protein powder, creatine, and anabolic steroids, is relatively difficult to obtain with accuracy (13), but this information is important to ascertain to raise awareness of the extent of harms associated with their use. Protein shakes and powders were reportedly used by 34.8% of high school adolescent boys in the US (14), and 25% of adolescent boys in Australia (15). When studies are conducted among specific athlete groups, a higher prevalence of use is found, with 42% using whey protein powders, 29% using protein bars, and 29% using pre-mixed protein drinks (4). The self-reported prevalence of creatine use among adolescents is around 5%, which is relatively low, considering that it is readily available (16). Steroid use is much less common among adolescent boys, with prevalence reported to range from 1.5% to 5.9% in large studies of Australian and US high school males (14,17,18). Large scale European prevalence data indicates that overall, 1% of adolescents have used anabolic steroids, but this was higher in some countries, including Bulgaria, where 7% of adolescent boys are using these substances (19). In light of the expansion of the supplements industries and consequent increased availability of these products, gaining current prevalence rates in adolescents is important.
As well as understanding usage rates, identifying factors that contribute to use of supplements provides greater opportunities to intervene and prevent harmful use. Two broad factors have been identified as being linked to supplement use -a desire to enhance appearance, or a desire to enhance performance (20). In addition to these factors, it is useful for researchers to fully understand the demographic, psychological, and contextual predictors of muscle building supplement use in order to more accurately prevent physical and psychological harm (21). Gender is commonly accepted as a predictor of muscle building supplement use in that males are much more likely to use supplements and steroids than females (22,23). A relationship between ethnicity and the use of muscle building supplements seems to exist, but analyses of those relationships are limited to ethnic groups located in the US (14).
In relation to appearance motivations for supplement use, research has found that adolescent boys who reported using muscle building supplements, including protein powders, were more likely to have higher levels of body dissatisfaction (24). These findings are also replicated in US samples where drive for muscularity was found to predict changes in performance enhancing substance use and weightlifting behaviour (25). In addition, common predictors of body dissatisfaction such as social appearance comparison, parent comments about appearance, and peer influences on appearance seemed to be particularly important in predicting the use of supplements in previous research (26).
Work with younger children (mean age ~9 years) also indicated that reductions in positive affect and pressure from the media, parents, and peers to be muscular predicted the use of muscle-building strategies 16-months later (27).
Given the advertised effects of supplements on enhancing sporting performance, sports participation is likely to be a relevant contributing factor to muscle building supplement use. Consistent with this contention, use of performance-enhancing substances is known to be more common among amateur and recreational athletes, for whom there are no penalties of supplementation (28,29). In relation to adolescents, research has found that those who participate in sports or are involved in sports teams were significantly more likely to use dietary supplements or to report use of protein powders for muscle enhancement than those who do not participate in sports (14, 23). However, the relationship between muscle building supplement use and participation in particular individual sports requires further research attention.
Given the broad access to supplements in Australia, it is important to understand the extent to which adolescent boys are using these substances, and the predictors of their use. This study aimed to explore the prevalence of protein supplement, creatine, and anabolic steroid use among Australian adolescent boys, and to determine the demographic, physical, and psychological factors that predict intentions to use and use of protein powder. It was hypothesised that after accounting for age, ethnicity, and BMI, lower body esteem and higher drive for muscularity would statistically predict greater intentions to use and actual use of protein powder. Furthermore, the addition of sports participation and weight training was expected to account for additional variance in both intentions to use and use of protein powder with greater sports participation and greater involvement in weight training being positively associated with intentions and with use.

Design
This research consists of secondary data analysis of baseline data from the control and intervention groups (successive yearly cohorts) of an evaluation of the impact of school-based steroid prevention program in adolescent boys (paper blinded for review) that was conducted in an all-boys school in Australia. High-risk human ethics approval was obtained from the (blinded for review) human ethics committee.

Participants
Students in two successive cohorts of grade 10 (aged 15-16 years) at an independent boy's school in Melbourne, Australia, were invited to participate in the research. Informed written consent was received from boys and their parents. A total of 237 participants, aged between 14 and 16, were included in the analysis.

Procedure
Participants completed the online questionnaire, delivered via the survey software Qualtrics, in Health and Physical Education class time, on their school-issued devices, while supervised in 'exam-like' conditions by a researcher and their class teacher.

Measures
The baseline questionnaire comprised standardized measures (see Table 1) and was used to collect data that would facilitate evaluation of the replication of the Athletes Learning and Training to Avoid Steroids [ATLAS] program. Therefore, the questionnaire included measures identical to those used in the original ATLAS evaluation research. Additional measures were used in the evaluation trial (Reference blinded for review) but are not reported here.

Data Analysis
All data was analysed using SPSS version 24. Two separate regression models were tested to assess for predictors of intention to use and actual protein powder use among participants. Regression models focused on protein powder due to the low proportion of participants using either creatine or anabolic steroids. First, a hierarchical linear regression was conducted to determine predictors of intention to use protein powder. Predictors entered into step one were demographic variables age, ethnicity, and BMI, followed by the addition of drive for muscularity (beliefs subscale) and body esteem (appearance subscale) at step two.
Step three included two additional variables of 'number of sports' participants engaged in and weight training (y/n). All categorical variables were dummy coded for entry into the models.
Second, a hierarchical logistic regression was conducted to assess predictors of protein powder use (yes/no). Similar to the first analysis, predictors initially entered into the model included demographic variables of age, ethnicity, and BMI at step one, followed by drive for muscularity and body esteem at step two. The number of sports participants engaged in and weight training were added in step three of the model.
Prior to regression analyses, the data was tested for assumptions of multicollinearity, normality, homoscedasticity as well as outliers and influential cases. The assumptions were sufficiently met with only slight deviations from normality and homoscedasticity observed. Given the large sample size and the absence of multi-collinearity and influential cases, the data was considered robust to small irregularities and the analysis was continued.
To control for multiple comparisons, an adjusted significance value of α=.025 was applied. Due to the nested models within the analysis, a standard Bonferonni method was employed for the calculation of the adjustment. This method enabled the recognition of the two regression models as two hypotheses rather than treating each variable across two models as a separate test. A-priori power analysis for multiple regressions revealed that for an anticipated medium effect size (Cohen's f 2 ) of 0.15, a desired statistical power level of 0.80 and 7 predictors with a probability level of 0.01, the minimum required sample size is 147, so our sample of 237 exceeds this number.

Sports Participation
Participants' self-reported engaging in a range of sports. The majority of participants listed participation in one sport (36.7%, n = 87), two sports (31.6%, n = 75), three (17.7%, n = 42), or four or more sports (13.1%, n = 31). Details of the type of sports that boys were engaged in are presented in Table 2. The sport with the highest proportion of participation was Australian Rules Football, followed by weight training and basketball. Participation in team and individual sports was relatively evenly represented.

Supplement Use
Self-reported supplement use and intentions to use supplements are presented in Table 3. The most commonly reported supplement used was protein powder, with a little under half of the sample reporting use. Half of participants reported use of at least one of protein powder, anabolic steroids, or creatine supplements (50.2%, n = 119). The proportion of participants who intended to use supplements was higher than the proportion who had actually used supplements for each of protein powder, creatine, and steroids.

Predicting Intentions and use of Protein Powder
Bivariate correlations (see Table 4) showed that muscularity beliefs, number of sports played, and engagement in weight training were all significantly associated with both intentions to use and actual use of protein powder, such that higher muscularity beliefs and greater engagement in sports and weight training were associated with higher intention to use and greater use of protein powder. Effect sizes were small to moderate. With the exception of age, for which a small inverse relationship with intention to use protein powder was revealed, no other independent variables were significantly correlated with protein powder intentions or use. Older age was also associated with lower muscularity beliefs and greater body esteem of small effect size. As expected, body esteem had small inverse correlations with muscularity beliefs and with body mass index. Muscularity beliefs was positively correlated with weight training, of small magnitude, but not correlated with number of sports played. Number of sports played and weight training were positively correlated at small effect size. The positive correlation between protein powder intentions and use was of large magnitude.
Predicting intentions to use protein powder. The summary statistics for the hierarchical linear regression analysis predicting intentions to use protein powder are presented in Table 5 Higher drive for muscularity, participation in a higher number of sports, and greater engagement in weight training, significantly predicted greater intent to use protein powder. When controlling for the effect of the number of sports, weight training remained the strongest contributor, accounting for 15.2% in the change in intentions.
Predictors of current protein powder use. Summary statistics for the hierarchical logistic regression predicting current protein powder use are presented in Table 6. The first step of the model, containing age, ethnicity, and BMI was not significant, X 2 (3) = 0.24, p = .972, and none of the independent variables significantly predicted protein powder use at this step. The inclusion of body esteem (appearance subscale) and drive for muscularity (beliefs) in the second step produced a nonsignificant model, X 2 (5) =7.20, p = .206. Only muscularity beliefs significantly predicted protein powder use, with stronger beliefs predicting greater likelihood of using protein powder. For the final model, the inclusion of number of sports and weight training produced a statistically significant model, X 2 (7) = 39.20, p < .001. Following adjustment for multiple comparisons (critical alpha cut-off set at .025), only number of sports played and weight training contributed significantly to the model.
The strongest predictor of use of protein powder was weight training. Boys who trained with weights were 3.8 times more likely to report use of protein powder than boys who did not. In addition, for every additional number of sports played, boys were 1.4 times more likely to use protein powders.

Discussion
In this paper, we have presented the prevalence and predictors of the use, and intentions to use protein powder, creatine, and anabolic steroids among Australian adolescent boys. We found that half of our participants had used protein powder, and that 62% intended to use protein powder. Use of creatine and anabolic steroids was lower than use of protein powder, but greater proportions of participants reported intentions to use these products. We found that the number of sports that boys participated in, and engagement in weightlifting, were the strongest statistical predictors of use, and intentions to use protein powder, more so than body esteem, drive for muscularity, age, ethnicity, or BMI. Our hypotheses were partially supported, in that sports participation and weight training, as well as drive for muscularity, were significant predictors of use, and intentions to use protein powder, but contrary to our hypotheses, body esteem was not.
The prevalence of the use of protein powder in this universal school sample is much higher than prevalence from data collected among a US sample 5 years ago. In the Project Eat study, 34.7% of adolescent boys (mean age 14 years) were using protein shakes and 5.9% reported using steroids (30). The prevalence found in our study was more closely aligned with an Australian sample from a similar time period, that found that 42% of adolescent athletes (13-18 years) had used whey protein powders, 29% consumed protein bars, and 29% used pre-mixed protein drinks (4). Replication with other universal samples will be required to determine whether supplement use is higher among Australian adolescent boys than adolescent boys from the US or elsewhere, or whether usage rates are increasing over time.
Our findings that sports participation and weight training were associated with greater use, and intentions to use protein powder is consistent with past research with adolescent boys which has also generally found that those who are playing sports are more likely to use muscle building substances.
However, the exact nature of the association is not yet clear. In the Project Eat study, adolescents participating in sports were significantly more likely to report muscle-enhancing behaviours, including the use of protein powder (14). In their systematic review of the factors predicting the use of doping and performance enhancing substances in young people, Nicholls and colleagues found that the relationship between sports participation and doping use was unclear. Of the five studies that were reviewed, some found no differences among athletes and non-athletes, higher use among athletes, or higher use among non-athletes (22). The review did, however, demonstrate that consistent with our findings, there was a clear relationship between engagement in strength-based sports and activities, for example weight training, American wrestling, and American football, and PIED use (22).
Our study findings are consistent with past literature that has reported greater supplement use in adolescents with body dissatisfaction (24). Our findings add greater specificity to past literature by demonstrating that muscularity beliefs, but not body esteem, were associated with higher intention to use, and actual use of protein powders. This distinction is important, as it demonstrates that it is the specific focus on concerns about muscularity, rather than general perceptions about appearance that are related to engagement in body change strategies in adolescent boys. This distinction provides a useful insight into potential targets for intervention, suggesting that improving body image in general, rather than focusing on muscularity specifically, would be likely to be less effective to address supplement use.
The supplements and doping literature is filled with examples of limitations to research due to the nature of the questions and scales used, and this research is no different. Given the time constraints, and limited attention span of adolescent boys, it is important that questionnaires are kept as short as possible, but this means that we miss out on layers of detail from not having the capacity to ask multiple questions in different ways. In this study, we asked direct questions about the use, and intentions to use protein powder, creatine, and anabolic steroids, but it is possible that there are better means of determining this information. Questions about sport participation could also have been more specific, in order to determine the level (amateur, recreational, elite, etc) and amount of the sport played (occasionally, one per week, etc). Finally, the numbers of boys consuming creatine and anabolic steroids were too low to allow us to complete regression analyses to predict the use of these products.
Our findings of high use of protein products in mid-adolescent boys have implications for the development and implementation of education, prevention, and intervention programs in a range of settings. First, this research justifies the inclusion of education about PIEDs and supplements in the school curriculum, aimed at a universal audience. In terms of timing, our findings of relatively high protein powder use among 14-16 year-old adolescents suggests that it is important that intervention and education programs to prevent supplement use take place when boys are 12-13 years old, before these attitudes and behaviours are established.
Our research found that adolescents who were engaged in weight lifting were particularly more likely (3.8 times) to be engaging in current protein powder use. This presents an opportunity to engage a potentially high-risk group of boys in selective intervention program to prevent current harm, and future PIEDs use. Education and professional development programs for teachers, personal trainers, gym and fitness centre staff, and coaches is also required in order to support programs targeted at boys. Whole-of-centre approaches towards making gyms, fitness centres, and other settings where weight lifting occurs more supportive of health (as opposed to appearance-related) motivations for physical activity are also advised.
In terms of the content of education programs, it is suggested that we look towards other evidencebased drug education programs that have been found to prevent the use of drugs and alcohol use among adolescents. Some school-based drug education programs designed for universal groups, regardless of their risk level, have produced small to moderate effects in terms of health behaviour change (31,32). Specific doping prevention interventions have also been effective, for example the Hercules program (33), and the Athletes Training and Learning to avoid Steroids [ATLAS] program (34), indicating potential directions for prevention programming.

Conclusions
We found that almost one in two adolescent boys aged 14-16 years was using protein powder, and that the use of this supplement was linked to participation in weight lifting, the number of sports that the boys were engaged, and drive for muscularity, although the sports related factors were more important in predicting intention and use than body image or demographic factors. This research provides important prevalence data that can inform clinical paediatric practice, as well as intervention and prevention programs.
Paediatricians should be aware of the high proportion of adolescent boys using protein powders and other muscle building supplements. Questioning patients about their weight training and muscle building supplement use is appropriate in order to determine whether boys are using supplements that might affect mood, growth, endocrine, and cardiovascular functioning. Advising boys who are currently lifting weights against using supplements with high levels of caffeine, and other ingredients is warranted.
There is very little attention to supplement education in the Australian Health and Physical Education Curriculum (35), and when supplement use is raised, in the optional Victorian Certificate of Education [VCE] Health and Physical Education syllabus, the focus is on doping in sport, and not a critique of the wider use of performance and image enhancing substances. We suggest addressing this gap in order to inform parents, teachers, and adolescents, about the potential impact of using these supplements, as well as creating specific indicated interventions for boys who are engaging in weight-lifting.

Competing interests
The authors declare that they have no competing interests to disclose.  Use of Protein Powder, creatine, and anabolic steroids Direct questioning items from ATLAS evaluation studies (32,33) includ "Have you ever used (protein powder/creatine/anabolic steroids)" and participants responded by selecting an option from 'I don't know what is', 'no, never', 'yes, once or twice', or 'yes, more than three times'.
Intentions to use protein powder, creatine, and anabolic steroids Items from ATLAS evaluation studies (32,33) asked participants to indicate the degree to which they agree or disagree (on a 7-point resp scale) with the statements 'I intend to try or use (protein powder/creatine/anabolic steroids)'.   Note. DFM = drive for muscularity *p<.05, **p<.01, ***p<.001