The present study is characterized as cross-sectional, consisting of four groups. Of these, three were composed of non-conventional professions (models, athletes, and ballet dancers) and a control group (undergraduate students). All groups were submitted to the same assessment protocols (Fig. 1).
The sample consisted of 41 female teenagers and young adults aged 14 to 24 years who belonged to modeling agencies, sports teams, dance schools and universities from the city of Porto Alegre, Rio Grande do Sul, Brazil. The subjects were allocated according to their daily occupation: models’ group (MG = 11), ballet dancers’ group (BG = 11), athletes’ group (AG = 8) and undergraduate students’ group (UG = 11). Modeling agencies, sports clubs and dance schools also signed an assent form. The study was approved by the Research Ethics Committee of the University (CAAE: 67847317.5.0000.5347), and was conducted according to the standards proposed by the Helsinki Declaration . The participants and their legal guardians (if participants are under 18 years old) were informed about the experimental protocol and the potential risks, providing written informed consent prior to participation (Fig. 1).
See Fig. 1.
Height (HEI) and body mass (BM) measurements were carried out. With these values, body mass index (BMI) was calculated (BMI = BM[kg]/HEI2[m]). In addition, bone mineral density (BDM), fat mass, fat-free mass (FFM) and fat percentage (%F) were evaluated in all participants (Lunar Densitometer model DPX.L, “Dual Energy X-ray Absorptiometry”). Data were processed by the system provided for each segment and the BDM results were expressed in g/cm2.
Cardiopulmonary exercise testing (CPET)
Peak oxygen consumption (VO2peak) was evaluated by open circuit ergospirometry during a maximum treadmill run test. The highest value reached in the last minute of exercise was used. According to the manufacturer’s instructions, the ergospirometry device (Quark CPET, Cosmed, Italy) was manually calibrated using the known concentrations of gases (21% O2 as reference concentrations, 12% O2 and 5.09% CO2 for calibration). The individuals had the possibility and time to adapt to the treadmill (Quinton Instruments, Seattle, USA) and the ergospirometry device. The participants performed a single progressive maximal exercise test, which consisted of walking for 4 min at 4 km h−1 followed by increases of 1 km h−1 at each minute until exhaustion. The individuals were verbally encouraged during the test to reach their maximal performance. To verify an exhausting effort, each participant had to meet at least one of the following criteria after the end of the test: (1) VO2 plateau that was defined as an increase in VO2 of less than 2.4 mL kg−1 min−1 with a corresponding increase in exercise intensity [30, 31], (2) obtainment of respiratory exchange ratio (RER) ≥ 1.0 [32, 33], (3) perception of effort greater than 17 (very intense—Borg’s scale of perceived exertion) . Heart rate was monitored using a heart rate monitor (FT1–POLAR). In addition, the individuals demonstrated evident signs of extreme physical effort at the end of the test, such as facial flushing, sweating, hyperpnea and unstable gait [32, 33]. According to these criteria, VO2peak of all participants were considered valid. All plots used in the determination of these points used breath by breath gross values. Two independent reviewers blindly determined VT and VCP following the criteria described above.
One-repetition maximum and muscular endurance tests
Maximal strength was obtained by one-repetition maximum (1RM) test for bilateral knee extensors (KE) and bilateral elbow flexors (EF). Bilateral elbow flexion exercise was performed with barbells and dumbbells, whereas bilateral knee flexion exercise was performed in a strength training equipment (Konnen Gym). The subjects were familiarized with all procedures in a single session. On the test day, the participants carried out a warm-up for five minutes in a cycle ergometer, after having performed specific movements for the exercises of the tests. The 1RM was defined as the heaviest weight a participant could lift once with a proper lifting technique, without compensatory movements, with no more than five attempts with five minutes of recovery between attempts. After each successful performance, the weight increased until a failed attempt occurred [35, 36]. The rate of the gradual increase in load was dependent on the participant’s self-perceived capacity and readjusted for the next series using Lombardi  calculations. Performance time for each contraction (concentric and eccentric) was 1.5 s controlled by an electronic metronome (Quartz, CA, USA) . After the performance of 1RM test, the load corresponding to 60% of 1RM of each individual was fixed. The subjects performed the maximum number of possible repetitions until fatigue. The number of repetitions performed was used as a measure of muscular endurance (ME) .
To determine the flexibility of the hip, back and posterior muscles of the lower limbs, the sit and reach test was performed using the Wells bench. The subject remained barefoot, in a sitting position facing the equipment with the soles of the feet flat on the bench, stretched forward along the bench’s metric demarcation, with outstretched arms and overlapping hands, seeking the greatest distance in three attempts, maintaining the value reached for 1 s. The highest value achieved was considered as a measure of flexibility .
For the performance of the biochemical analyses of the study, the subjects should be fasting for eight hours and performed the assessments between 48 and 72 h after the last exercise session . Initially, a 30-min rest was performed, for further collection of 8 ml of a vein in the antecubital region. The samples were stored (vacuitaner tubes with EDTA) and centrifuged (3.500 rpm for 10 min), the plasma was aliquoted and frozen (-80°) for further analyses of fasting glucose (FG), fasting insulin (FI), total cholesterol (TC), triglycerides (TG), C-reactive protein (CRP) and high-density lipoprotein (HDL). Glucose levels were analyzed by enzymatic colorimetric method (Cobas C111, Roche, Diagnostics, Basel, Switzerland). Low density lipoprotein (LDL) was calculated by Friedewald formula (LDL:TC-HDL-TG/5) . Plasma concentrations of FI were evaluated with commercial kits for humans (DRG International, Springfield, USA) by enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s instructions. For the lipid variables, the Brazilian Guidelines of dyslipidemias and atherosclerosis prevention  were used with determined health risk values for TC (< 170 mg/dL), HDL (> 45 mg/dL), LDL (< 110 mg/dL) and TG (< 90 mg/dL). Values lower than 100 mg/dL and between 1.9 and 23.0µUI/mL were used for FG and FI, respectively . For the inflammatory profile, the value of <1 mg/L was used to determined values for CRP .
To evaluate dietary consumption and diet quality, a three-day dietary record was applied, describing daily dietary intake with the respective quantities. The record procedure was carried out as follows: each participant recorded all foods and beverages ingested in three days (two typical days and one atypical day of the weekend), also describing the eating schedule, quantities and when possible, the brand of the food. After filling in, all information was checked by a trained researcher, so that there would be no doubt about what was described. The dietary records were further calculated with the DietWin Professional Nutrition Software (Brubins CAS, Brazil).
Depressive symptoms assessment was carried out through the transcultural and self-applied CES-D depression scale, consisting of 20 statements on a Likert scale with four possible answers corresponding to number 1 (rarely or never – less than 1 day), 2 (few times – 1 to 2 days), 3 (a considerable time – 3 to 4 days) and 4 (all the time – 5 to 7 days). Four items of the instrument are presented with positive sense (blocking the trend to repetitive responses), such items are reversely scored (higher scores indicate greater amount of depressive symptoms). The cutoff score for Brazilian adolescents and young adults is 15, in which individuals with this score or higher are considered at risk for depression .
Normality of data distribution was evaluated by Shapiro Wilk test and homogeneity of variables was assessed by Levene test. Mean and standard deviation values were used for descriptive purposes. ANOVA one-way was used to analyze the differences between groups. LSD post hoc was used to establish the location of the differences between groups. Partial eta squared (η2) was calculated as a measure of effect size (ES). Values of 0.01, 0.06 and above 0.15 were considered small, medium, and high, respectively . Kruskal–Wallis test was applied for the variables that did not present normal distribution. When the Kruskal–Wallis test was significant, pairwise comparisons were performed by Mann–Whitney U test for independent samples. Statistical analyses were performed with SPSS software (version 20.0, SPSS, Inc., IBM Company; NY, USA). The significance level adopted was p < 0.05.
Of the total sample, AG presented the youngest age (years) between groups (16.38 ± 3.02, p = 0.002), followed by BG (18.09 ± 3.21), MG (20.82 ± 2.52) and UG (21.91 ± 2.30). For career duration (months), the highest values were presented by BG (147.27 ± 65.97), followed by AG (132 ± 44.44) and MG (56.73 ± 36.04) (p = 0.001). Of the four groups, only three individuals of the AG had not yet gone through the first menarche (p = 0.004). For nutritional and psychological counseling, UG presented 5 individuals with nutritional counseling and 7 with psychological counseling, MG had 7 and 4, BG had 2 and 5 and AG presented 6 and 6, respectively.
Statistically significant differences between groups were found for height, BMI, WHtR, %Fat, fat mass, 1RM of KE and EF and flexibility (p < 0.05). In BMI, MG presented the lowest values (18.1 ± 1.37, p = 0.018) when compared to the other groups, however, it was within the expected for sex and age . For WHtR, all groups presented lower values than the cutoff value for cardiovascular health . In the 1RM of KE and EF, BG presented the lowest values in both limbs (1RM KE: 68.00 ± 16.89, p = 0.031; 1RM EF: 13.91 ± 2.39 p = 0.001). On the other hand, in the values obtained for flexibility, AG (50.38 ± 5.71 cm) and BG (45.77 ± 6.71 cm) presented significantly greater values compared to the other groups (UG: 36.41 ± 7.73; MG: 31.32 ± 9.96). For BMI, 14 individuals (34%) presented results below the normative values  for sex and age. As for flexibility, only 4 (10%) girls presented values considered bad  for sex and age. Even without statistically significant values, for BDM, 26 individuals (63%) showed results below the normative values [51, 52]. The results regarding body composition, cardiorespiratory fitness, maximal dynamic strength, muscular endurance and flexibility are presented in Table 1, whereas the individual responses are presented in Table 4.
When analyzing the lipid (HDL, LDL, TG, CT) inflammatory (CRP) and glycemic (FG and FI) variables, no statistically significant differences were found between groups (p > 0.05). The effect size values found were HDL: η2 = 0.14, LDL: η2 = 0.05, TC: η2 = 0.02, TG: η2 = 0.09, FG: η2 = 0.12, FI: η2 = 0.07, and CRP: η2 = 0.07 (Fig. 2). The individual responses are presented in Table 4. For HDL, only one individual (2% of the total) presented a non-normative value (> 45 mg/dL), for TC, 6 individuals (15%, < 170md/dL), for TG, 9 individuals (22%, < 90 mg/dL), for FI, 9 individuals (22%, from 1.9 to 230µUI/mL), and for CRP, 13 individuals (31%, < 1 mg/dL).
When analyzing the dietary control, no statistically significant differences were found between groups (p > 0.05). The mean energy intake (kcal) presented by AG is below the recommended (individually) for athletes who are exposed to the training loads described in this study (Table 2).
The results regarding the depressive symptoms score are presented in Table 3. No statistically significant differences were found between groups (p > 0.05). All groups presented values above the cutoff points (> 15) for mental health. In addition, 53.6% of the girls were under psychological counseling. The individual responses are presented in Table 4. Of the total subjects of the study (41 individuals), 100% presented values that evidence a trend to depressive symptoms.