Elite gymnasts show rising irisin during eight-week pre-competition training

Circulating irisin, a muscle-derived hormone associated with responses to exercise, rose during an eight-week pre-competition training period in a small pilot study of eight nationally classified female rhythmic gymnasts. Mean serum concentration increased from 1.357 nanograms per milliliter at the beginning of the program to 4.597 ng/mL in its final week. The mean within-athlete rise was 3.240 ng/mL, with a 95% confidence interval of 0.338 to 6.143 ng/mL and a standardized effect size of 0.93. The difference was significant at a nominal level, but the observational design, eight-person sample and measurement uncertainties mean that the result cannot show that training caused the increase or that irisin improved performance, recovery or psychological well-being.

The longitudinal project followed the eight gymnasts on a single high-level team in China through an eight-week pre-competition mesocycle. Their mean age was 19 years, mean training experience was 11.75 years, and mean height and weight were 167.45 centimeters and 49.66 kilograms, respectively. The program called for training on six days each week, with morning and afternoon sessions of approximately four hours on most days. It included flexibility work, apparatus-specific technical preparation, jumps and balance sequences, routine repetitions, twice-weekly strength and conditioning, once-weekly psychological skills training and Sunday recovery. The full available squad was included, so no formal sample-size calculation was performed. All eight athletes completed the monitoring period without missing observations in the variables analyzed, but a single-team cohort cannot represent rhythmic gymnasts generally.

The researchers combined subjective and objective measures rather than relying on a single marker of training stress. Daily rating of perceived exertion, or RPE, was recorded twice on each monitored day, while the Chinese-language Athlete Training State Monitoring Scale assessed eight dimensions: emotional stress, general well-being, fatigue, self-regulation, physical recovery, psychological exhaustion, psychological fatigue and self-efficacy. Heart rate was monitored during Weeks 1, 3, 5, 7 and 8, blood samples were analyzed in Weeks 1 and 8, urine was sampled before morning training and after afternoon training, and body composition was measured after training in Weeks 1, 5 and 8. The design captured repeated responses but did not include a control group, so it could describe changes associated with preparation without establishing their causes.

Perceived exertion changed over the monitored weeks, whereas the eight psychological subscales did not show clear evidence of longitudinal change after statistical correction. Weekly mean values differed across weeks for both the first and second daily RPE ratings, with partial eta-squared values of 0.655 and 0.641, respectively, and nominal p values below 0.001. By contrast, all psychological week-effect p values were at least 0.1211. Heart-rate exposure also varied. Absolute time in the 50%–59%, 60%–69% and 70%–79% zones of age-predicted maximum heart rate changed across weeks, with nominal p values below 0.001, but after exposure was normalized to each athlete’s total valid monitoring time, only the 50%–59% zone retained a clear week effect. Thus, some apparent changes in minutes reflected differences in how much daytime activity was recorded rather than a systematic shift in relative intensity.

Several blood and urine measures changed, but their interpretation was mixed. LDL cholesterol and apolipoprotein B were higher at Week 8, with nominal p values of 0.028 and 0.032, while total cholesterol and lipoprotein(a) were lower, with nominal p values of 0.003 and 0.045. The investigators cautioned that this discordant lipid pattern should not be treated as evidence of a specific training-induced adaptation, particularly given inter-individual variability, biological variation and analytical variation. Mean corpuscular hemoglobin concentration remained below its reference range at both blood assessments, while mean corpuscular volume decreased despite remaining within its conventional reference interval. The study did not measure ferritin, transferrin saturation, dietary intake, endocrine markers or menstrual-cycle status, so it could not determine whether the low concentration reflected iron deficiency or low energy availability.

Urine protein showed evidence of variation across weeks and was higher in post-training samples overall, but the week-by-sampling-condition interaction was not significant. Urobilinogen changed over the study period, and its pre- versus post-training pattern differed by week, while urine pH, specific gravity and leukocytes showed no clear evidence of the tested effects. The pre-training and post-training samples represented the beginning and end of an entire training day, not measurements around one identical session. Hydration, diet, sampling timing and the semi-quantitative nature of some test-strip results could influence the findings, so the results did not establish a change in hydration status. Ketone results were retained for description only because the records could not reliably distinguish an unrecorded result from a confirmed negative.

The study also examined whether changes in irisin were associated with psychological measures. An athlete’s increase in irisin from Week 1 to Week 8 was strongly negatively associated with their mean emotional-stress score across the five monitoring assessments, with Pearson’s r equal to −0.859, a 95% confidence interval of −0.974 to −0.390 and a nominal p value of 0.006. The change was also positively associated with mean psychological exhaustion, although that estimate was less certain, with r equal to 0.704, a confidence interval extending slightly below zero and a nominal p value of 0.051. Correlations with mean RPE, general well-being, fatigue, self-efficacy, self-regulation, physical recovery and psychological fatigue did not provide clear evidence of association. These are exploratory associations among only eight athletes, not evidence that irisin caused any psychological response.

Sensitivity analyses showed why the irisin-stress result requires restraint. Mean emotional stress was positively associated with baseline irisin, although the confidence interval included zero, and negatively associated with Week 8 irisin. Baseline irisin was also inversely associated with the subsequent change, while the two athletes whose irisin decreased had the two highest starting concentrations. Mathematical coupling between baseline and change scores, regression to the mean, measurement variability and small-sample bias could therefore contribute to the pattern. Moreover, the main correlations linked the change in irisin with each athlete’s average psychological score, not with psychological change over time. The results do not demonstrate that irisin buffered emotional stress or that rising irisin reflected better psychological adaptation.

Measurement limitations add another layer of uncertainty. The irisin concentrations were measured with a commercial enzyme-linked immunosorbent assay, and duplicate optical-density readings consistently differed by position, so a technical effect could not be excluded. The lower calibration range also showed discrepancies when reconstructed from the assay equation, and several early samples fell in that range, potentially influencing the estimated size of the Week 1-to-Week 8 increase. Bioelectrical impedance estimates of body fat and visceral fat may likewise be affected by hydration. External workload was documented broadly, but apparatus repetitions, jump counts, landing counts and direct performance outcomes were not systematically recorded. Maximum heart rate was estimated from age rather than measured in a maximal test, adding possible error to zone classification.

Taken together, the pilot study supports the value of examining several domains at once while emphasizing how little can be concluded from eight athletes over one preparation block. Irisin increased markedly, perceived exertion and selected physiological markers varied, and a strong association appeared between irisin change and average emotional stress, but the broader psychological scales did not show clear longitudinal change. The findings are best treated as hypothesis-generating: the team was observational, numerous outcomes were tested without an overall correction for multiplicity, and important influences such as nutrition, sleep, menstrual cycle, hydration and stress outside sport were not controlled. Larger studies with comparison groups, more repeated measurements, quantified workload, direct performance outcomes, nutritional and endocrine assessment and validated irisin methods will be needed to determine whether multidimensional monitoring can support individualized training and recovery decisions in elite rhythmic gymnastics.

Subject of Research: Longitudinal monitoring of physiological, psychological and irisin responses in elite rhythmic gymnasts during pre-competition training

Article Title: Exploratory longitudinal monitoring of physiological, hematological, psychological, and irisin responses during an 8-week pre-competition period in elite rhythmic gymnasts

Article References: Ye, M., Qian, Y., Ren, S., Kéringer, J., Lin, C., Wang, M., & Radák, Z. (2026). Exploratory longitudinal monitoring of physiological, hematological, psychological, and irisin responses during an 8-week pre-competition period in elite rhythmic gymnasts. Frontiers in Sports and Active Living, 8, Article 1925319. https://doi.org/10.3389/fspor.2026.1925319

Image Credits: AI Generated

DOI: 10.3389/fspor.2026.1925319

Keywords: rhythmic gymnastics, elite athletes, irisin, training-load monitoring, psychological stress, pre-competition training, heart-rate zones, blood biomarkers, physiological monitoring, Exploratory, longitudinal, monitoring