This cross-sectional study consisted of a questionnaire sent via the Web and based on volunteer participation.

The research team elaborated the questionnaire and then translated its original English version into six languages (Hungarian, Italian, Japanese, Lithuanian, Portuguese, and Spanish). Subsequently, the questionnaire was translated back into English for the establishment of semantic and conceptual equivalence by the research group. The Qualtrics online research platform (Qualtrics, Provo, UT, 2020) was used for data collection. The survey was disseminated via the Web and using a snowball sampling method, in which participants were invited to fill in the questionnaire and share it with their friends and relatives. Dissemination was also implemented through posts on social media such as Facebook, Twitter, Instagram, LinkedIn, and WhatsApp. Inclusion criteria were (i) age between 18 and 65, (ii) use of at least one IPED, and (ii) consent to participate in the study. The data obtained were securely stored on a password-protected computer at the University of Hertfordshire, Hatfield, United Kingdom (United Kingdom).

Data collection took place during April and May 2020, which was the peak period of lockdown in all nations taking part in the study.

The survey comprised: (i) sociodemographic questions; (ii) questions on IPEDs use; and (iii) validated psychometric instruments, including the Exercise Addiction Inventory (EAI), Appearance Anxiety Inventory (AAI), and Self-Compassion Scale (SCS- Short Form).

The EAI (34) is a tool developed to measure addictive tendencies in exercise behavior. It includes six statements gaging the classic symptoms of addictions (i.e., salience, mood modification, tolerance, withdrawal symptoms, social conflict, and relapse), which are rated on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The maximum score obtainable on the EAI is 30. A person scoring 24 or higher may be at risk of exercise addiction. This suggested cut-off score represents the top 15% of the total scale’s score. The EAI is presented as a valid and reliable psychometric instrument, used in many studies across various countries (34–36). Cronbach’s alpha in the present sample was 0.641, suggesting an acceptable internal consistency.

The AAI (36, 37) is a 10-item tool that measures cognitive and behavioral dimensions of appearance anxiety and symptoms associated with Body Dimorphic Disorder (BDD) (36). In this study, it is rated on a 4-point Likert scale that ranges from 1 (not at all) to 4 (all the time). Values of 21 or more correspond to the top 15% of the total scale’s score. Cronbach’s alpha in the present sample was 0.876, suggesting a very good internal consistency.

The SCS-Short Form (38) consists of 12 items and is related to self-compassion and emotional self-regulation. It comprises six subscales: (1) self-kindness, (2) self-judgment, (3) common humanity, (4) isolation, (5) mindfulness, and (6) over-identification. All items are rated on a 5-point Likert scale, from 1 (almost never) to 5 (almost always). A higher score suggests greater self-compassion. Cronbach’s alpha in the present sample was 0.852, suggesting a very good internal consistency.

IPED questions. Participants were asked about the forms of supplements or products they used to reach their fitness or physical appearance goals during self-isolation. Further, they were asked about their habitual mode of purchase: pharmacies, specialized shops, or the Internet (web shops). For comparison purposes, the IPEDs listed were the same as those in previous research conducted by Corazza et al. (17). This list was developed in consultation with medical doctors and sports dieticians.

The lead researchers performed an initial data familiarization to generate two IPED categories within the dataset: (1) “non-harmful IPEDs” and (2) “potentially risky (or hazardous) IPEDs.” The extant literature was consulted to refine the two categories and apply them to the full sample (22). The first category includes vitamins, minerals, proteins, amino acids, and natural extracts. The second consists of any other IPED that does not fit in the “non-harmful” category, such as medications, steroids, and stimulants. The list of the IPEDs in the two categories is shown in Table 1.

The Ethics Committee of the University of Hertfordshire, United Kingdom, approved the study (permission: HSK/SF/UH/00104). Also, ethical clearance was obtained from the Ethics Committees of the participating institutions from the various nations. The European General Data Protection Regulation, as well as the norms of the Helsinki Declaration, were rigorously followed during the work (39).

The IBM Statistical Package for Social Sciences (SPSS), developed for the Windows platform, was the software used for data analyses (IBM SPSS Inc., Chicago, Illinois). Basic descriptive statistics consisted of mean, standard deviation, skewness and kurtosis. Only hourly values in the question on ‘time spent online’ were not normally distributed. These values were expressed as median (IQR). The assumptions of normality were not violated for the rest of the analyzed data, which were expressed as means, standard deviations, frequencies, or percentages, as appropriate. Subsamples were compared with Student’s t-tests and Mann–Whitney U tests for the continuous measures. For categorical data, Chi-square tests and Fisher exact tests were employed. The level of statistical significance was set at alpha (α) = 0.05.

Based on previous works (15, 18) effect size was estimated to be d = 0.25. Thus, with an effect of 0.25, power set at 90% and an alpha level of 0.05, the total sample size was calculated to be 735 participants (allocation rate: N2/N1 = 1.8).

There were 736 IPEDs users from Brazil (n = 337; 45.7%), Italy (n = 134; 18.2%), Spain (n = 42; 5.7%), Lithuania (n = 78; 10.5%), Portugal (n = 51; 6.9%), the United Kingdom (n = 36; 4.9%), Japan (n = 31; 4.2%), and Hungary (n = 27; 3.6%). Their mean age was 33.05 (±SD = 10.06) years, and most were women (n = 473; 64.2%).

Among IPEDs users, 6.8% (n = 50) were susceptible to exercise addiction (EAI >24). Another 27.6% (n = 203) presented high levels of appearance anxiety (AAI ≥ 21), and another 24.9% (n = 183) revealed low emotional regulation-related self-compassion.

One or more IPEDs in the “potentially risky” category was used by 488 respondents (66.3%). The remaining participants (n = 248, 33.7%) used only IPEDs in the “non-harmful” category. Comparisons between users of “potentially risky” and “non-harmful” IPEDs are detailed in Table 2. Users of potentially risky IPEDs were younger (p < 0.001) and male participants (p < 0.001). They exhibited more elevated scores both on the EAI (p < 0.001) and on the AAI (p < 0.001). As to cross-cultural comparisons, in Brazil there was a significantly higher use of potentially risky IPEDs (p < 0.001), while in Japan the use of non-harmful IPEDs was prevalent (p = 0.004). Users of potentially risky IPEDs reported a higher incidence of both psychiatric disorders in general (p < 0.001) and of mood disorders other than depression, such as dysthymia, cyclothymia or bipolar disorder (p = 0.014). In the group of users of potentially risky (hazardous) IPEDs, there were statistically significantly more smokers (p = 0.012). Among those starting the use of IPEDs during the pandemic (n = 112; 33.1%), the largest part consumed only non-harmful products (p = 0.004). However, among those using IPEDs both before and during the pandemic (n = 524; 71.2%), most used potentially risky ones (p = 0.003).

Participants were asked about their prevalent source of IPEDs purchase. Most participants (n = 409, 55.6%) purchased IPEDs through pharmacies, shops, or specialized shops; 304 (41.3%) purchased IPEDs on the Internet, either on legal sites or on the black market, and 23 participants (3.1%) purchased IPEDs through other modalities.

Participants were compared based on the modality of IPEDs purchase (see Table 3 for detailed information). Participants who purchased IPEDs online were mostly men (p < 0.001) and had more elevated EAI scores than did their non-Internet purchasing counterparts (p = 0.004). Regarding cross-cultural differences, in Spain and Brazil there was a significantly higher percentage of participants purchasing IPEDs from “pharmacies/ shops/specialized shops. “Potentially risky” IPEDs were purchased online in greater proportions than they were in pharmacies/shops (70.1% vs. 63.3%), in contrast with “non-harmful” IPEDs, which were purchased more frequently in pharmacies or shops (36.7%) than on the Internet (29.9%). The difference, however, failed to reach statistical significance (p = 0.060). Although more users have consulted a health professional about the use of IPEDs when buying from pharmacies/shops (46.2%) than from the Internet (40.3%), the difference was not statistically significant (p = 0.114). The majority of users consumed IPEDs both before and during the pandemic (n = 504, 68.5%). Among those using IPEDs before but not during the pandemic, the majority purchased them from “pharmacies/shops/specialized shops” (p = 0.017).

Some significant differences emerged between individuals who buy IPEDs online versus those who buy them in shops. More men tended to buy these substances on the Internet rather than in specialized shops (43.8% vs. 30.1%; p < 0.001). Conversely, more women tended to buy IPEDs in shops rather than on the Internet (69.9% vs. 56.3%). Cross-country comparisons showed that a significantly greater percentage of individuals who preferred to buy IPEDs on the Internet was from the United Kingdom. Conversely, in Spain and Brazil, a significantly greater proportion of users reported purchasing IPEDs from pharmacies or shops. This result is new when compared to previous literature (17, 22), where respondents in the United Kingdom sample predominantly purchased IPEDs also from shops or pharmacies. This practice can be explained at times of lockdown involving social distancing. Still, it could signal a changing trend, which deserves more empirical attention in future works.

Internet buyers scored significantly higher on the EAI, suggesting a higher risk of developing exercise addiction. Recent literature suggests that excessive exercising is positively related to the unsupervised consumption of IPEDs (15, 18), which can be endorsed and facilitated by the online purchase of these substances. Online commercials and false adverts could increase the use of these drugs via misleading marketing strategies that promise physical and mental improvement by promoting them as alternatives to controlled medical or pharmaceutical products (43–45).

The majority of IPEDs users (66.3%) tended to consume potentially risky substances. This group of users of potentially risky IPEDs were mainly younger men who scored higher on the EAI and the AAI. There was no significant association between self-compassion, education level, or time spent on the Internet (both before and during the pandemic), and the use of risky IPEDs.

The group of consumers of potentially risky IPEDs also had a significantly higher incidence of smoking habits and of psychiatric disorders in general (and mood disorders other than depression in particular), when compared with the non-harmful substances group. The most frequently represented psychiatric disorders in the group of users of potentially risky IPEDs (which were the same as in the group of non-risky substance users) were anxiety (82.1%), depression (34.5%), and eating disorders (21.0%). An interesting finding was that individuals who were already using IPEDs before the COVID-19 pandemic, and continued doing so during the pandemic, used significantly more potentially risky substances when compared with both those who stopped taking IPEDs and those who started taking IPEDs during the pandemic. The new users of IPEDs during the pandemic predominantly preferred substances in the “non-risky” category, suggesting that those who initially approach these substances tend to prefer those considered (and advertised) as “safer.” Moreover, they might have started taking supplements during the pandemic as a way to reinforce their immune system against the virus. More individuals taking non-dangerous substances saw a doctor about the consumption of these products (48.0%) than did those taking potentially dangerous substances (41.3%), although the difference was statistically non-significant (p = 0.083). Nevertheless, even though negative effects connected with the use of supplements such as creatine, caffeine, and steroids have been broadly documented in the literature (46, 47), the newest IPEDs remain marginally studied and regulated (20, 48). As observed in other studies (49), the consumption of supplements among people who consistently use them is generally perceived as safe, acceptable, and needed for achieving the ideal body form, or weight, and fitness objectives.

This study presents new data on the subject, but also has some limitations. First, the questionnaire was disseminated online and contained self-reported measures, without any biological tests to confirm the data collected on substance use. The second limitation is that it is based on a non-stratified sample of volunteers, which may result in selection bias and, consequently, this sample may not represent the population. For example, it is possible that, because the questionnaire was accessed online by respondents who use the Internet, Internet buyers of IPEDs are overrepresented in the sample (e.g., in the United Kingdom, comparing to previous studies on IPEDs use). Also, the sample sizes by country are quite different, with the Brazilian sample being the largest. Third, within the two groups of (“potentially risky” and “non-harmful”) IPEDs, categories should be broken down further for a better understanding of the tendencies found regarding purchasing risky IPEDs online, and (not) seeing doctors about IPEDs use.