Associations Between Endocrine Status and Stress, Mood and Psychosomatic Status in Elite Handball Players

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for the opportunity to prepare and conduct a peer review of an article for an MDPI journal. The title of the scientific article was  Associations between endocrine status and stress, mood, and psychosomatic status in elite handball players, the aim of which was to investigate the relationship between endocrine profile (cortisol, testosterone, estradiol) and indicators of stress and mood in elite handball players. The added value of the article lies in the number of participants (n=584). However, upon analysis of the scientific article in its current state, it contains certain methodological gaps that require supplementation or extensive commentary. Below, I provide a point-by-point breakdown, organized by sections of the scientific article, of what needs to be corrected:

1. Abstract. The abstract is correctly structured and includes a clear objective, a description of the methods, and the main results. In this section, 1–2 sentences should be added regarding the study’s limitations (such as the cross-sectional nature of the study or the lack of strict control over wake-up time, which can be a key factor when it comes to cortisol). 
   
   
2. Introduction. The introduction to the research problem is very well described. The authors used 19 references in the introduction (19 citations in this section of the scientific article). However, it might be worth adding information in the introduction about why psychosomatic symptoms may correlate more strongly with biological markers (hormones) than questionnaire-based mood or stress (as later demonstrated in the results). Please respond to this suggestion. 
   
   
3. Materials and Methods. The authors used validated tools in their study (ELISA, PSS, POMS, and HBSC-SCL questionnaires). Furthermore, the study’s objective was clearly stated. No comments. However, I have a question for the authors of this scientific publication. Would it not be worthwhile to include information on how the phase of the cycle was verified (was it based solely on the questionnaire and the athletes’ self-reports, or on body temperature measurements/apps?) 
   
   
4. Results. The results are described clearly. The use of GLM models is a correct statistical choice. Although the use of partial eta squared was stated in the methods section, these values (or other effect size indicators) are missing directly in the results tables. 
   
   
5. Discussion and Conclusion. The discussion must absolutely be expanded to include a Limitations section. This section should address: the cross-sectional nature of the study (the lack of repeated measurements throughout the season for the same athletes prevents a cause-and-effect assessment) and the aforementioned issue regarding the timing of saliva collection for cortisol. 
   
   
6. References. Please add an information why did you used citation from 1971 year? McNair, D.; Lorr, M.; Doppleman, L. POMS Manual for the Profile of Mood States. San Diego, CA: Educational and Industrial Testing Service 1971. Or Cohen, S.; Kamarck, T.; Mermelstein, R. A global measure of perceived stress. J Health Soc. Behav. 1983, 24, 385-396.?

The article is scientifically valuable and addresses an important topic, but it requires significant revision.

Author Response

General comment: Thank you for the opportunity to prepare and conduct a peer review of an article for an MDPI journal. The title of the scientific article was Associations between endocrine status and stress, mood, and psychosomatic status in elite handball players, the aim of which was to investigate the relationship between endocrine profile (cortisol, testosterone, estradiol) and indicators of stress and mood in elite handball players. The added value of the article lies in the number of participants (n=584). However, upon analysis of the scientific article in its current state, it contains certain methodological gaps that require supplementation or extensive commentary. Below, I provide a point-by-point breakdown, organized by sections of the scientific article, of what needs to be corrected:

Response: The Authors would like to express sincere thanks to the Reviewer for careful reading and suggestion for improvement in the paper.

 

Comment 1: Abstract. The abstract is correctly structured and includes a clear objective, a description of the methods, and the main results. In this section, 1–2 sentences should be added regarding the study’s limitations (such as the cross-sectional nature of the study or the lack of strict control over wake-up time, which can be a key factor when it comes to cortisol).

Response: The Authors thank this comment of the Reviewer, the Abstract was completed with the suggested limitations as follows:

“Methods: In a cross-sectional study, salivary cortisol (with no strict control over wake-up time), testosterone, and - in female athletes - 17‑β‑estradiol concentrations were assessed in 584 elite handball players aged 14-35 years using ELISA.”

 

Comment 2: Introduction. The introduction to the research problem is very well described. The authors used 19 references in the introduction (19 citations in this section of the scientific article). However, it might be worth adding information in the introduction about why psychosomatic symptoms may correlate more strongly with biological markers (hormones) than questionnaire-based mood or stress (as later demonstrated in the results). Please respond to this suggestion.

Response: The Authors thank this recommendation of the Reviewer. We agree that psychosomatic symptoms may correlate more strongly with biological markers than questionnaire-based measures of mood or stress status. The biological explanation to this relation could be that they reflect somatic manifestations of physiological stress responses, whereas self-reported scales are influenced by subjective perception, reporting bias and psychological interpretation. We searched the scientific literature for evidence, study results and published confirmation of this assumption, but could not find relevant publications for this relationship. Therefore, we completed the Introduction section by assuming this relationship:

“Based on these considerations, the aim of the present study was to examine the associations between endocrine status and psychological (perceived stress and mood) and psychosomatic indicators in elite handball players. We hypothesized that psycho-somatic symptoms would show stronger associations with hormonal parameters than perceived stress or mood disturbance, reflecting the distinction between perceived stress and underlying biological activation processes.”

And the Discussion section was completed with the following addition in this respect:

“Psychosomatic symptoms, by integrating bodily and perceptual dimensions of stress, may therefore offer a more functionally relevant reflection of an athlete’s physiological state. The results of the present study support this hypothesis, as psychosomatic indicators demonstrated stronger associations with endocrine status than questionnaire-based measures of perceived stress and mood.”

 

Comment 3: Materials and Methods. The authors used validated tools in their study (ELISA, PSS, POMS, and HBSC-SCL questionnaires). Furthermore, the study’s objective was clearly stated. No comments. However, I have a question for the authors of this scientific publication. Would it not be worthwhile to include information on how the phase of the cycle was verified (was it based solely on the questionnaire and the athletes’ self-reports, or on body temperature measurements/apps?)

Response: The Authors thank this comment, as well. The Methods section on phase of the cycle was verification was completed as follows:

“The determination of menstrual cycle phase was based on self-reported data obtained through personal interviews with the athletes, including information on cycle length, bleeding patterns, regularity and the date of the last menstruation.”

 

Comment 4: Results. The results are described clearly. The use of GLM models is a correct statistical choice. Although the use of partial eta squared was stated in the methods section, these values (or other effect size indicators) are missing directly in the results tables.

Response: We thank this comment of the Reviewer. The manuscript was completed with the effect sizes in the text of the Results section.

 

Comment 5: Discussion and Conclusion. The discussion must absolutely be expanded to include a Limitations section. This section should address: the cross-sectional nature of the study (the lack of repeated measurements throughout the season for the same athletes prevents a cause-and-effect assessment) and the aforementioned issue regarding the timing of saliva collection for cortisol.

Response: The Limitations section follows the Discussion section, it was corrected by following the Reviewer’s comment as follows:

“The present study has several limitations. First, the cross‑sectional design precludes conclusions regarding causal relationships between psychological indicators and hormonal parameters. Second, although saliva sampling provides a practical and non‑invasive method for assessing endocrine status, variability related to sampling conditions - particularly the lack of strict control over wake‑up time, which is a key determinant of cortisol levels due to its pronounced circadian rhythm - may have affected the measurements.”

The Discussion section was completed as follows:

“In male athletes, higher levels of psychosomatic symptoms were associated with elevated cortisol levels, while in female athletes, associations were observed with testosterone levels (these results should be interpreted with caution, as cortisol levels may have been influenced by variability in sampling conditions, particularly the lack of strict control over wake‑up time).”

 

Comment 6: References. Please add an information why did you used citation from 1971 year? McNair, D.; Lorr, M.; Doppleman, L. POMS Manual for the Profile of Mood States. San Diego, CA: Educational and Industrial Testing Service 1971. Or Cohen, S.; Kamarck, T.; Mermelstein, R. A global measure of perceived stress. J Health Soc. Behav. 1983, 24, 385-396.? The article is scientifically valuable and addresses an important topic, but it requires significant revision.

Response: The Authors thank this suggestion, the References were corrected with the revised versions of the references mentioned by the Reviewer as follows:

21. Goulet-Pelletier, J.C.; Cousineau, D. A review of effect sizes and their confidence intervals, Part I: The Cohen’s d family. Quant. Meth. Psych. 2028, 14(4), 242-265. https://doi.org/10.20982/tqmp.14.4.p242
22. McNair, D.M.; Lorr, M.; Droppleman, L.F. EdITS Manual for the Profile of Mood States (POMS). Educational and industrial testing service. San Diego, CA: Educational and Industrial Testing Service 1992.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

General comments:

The manuscript investigates the associations between endocrine status (salivary cortisol, testosterone and estradiol) and psychological/psychosomatic indicators in elite handball players. The topic is relevant for athlete monitoring and the prevention of maladaptation to training, and the study benefits from a large sample of elite athletes across a wide age range.

 

Overall, the manuscript is generally well structured and addresses an important applied question. The sample size is a major strength and the inclusion of both endocrine and psychosomatic variables is potentially valuable. However, several methodological and statistical issues require clarification. In particular, the rationale for some analytical decisions, the interpretation of the findings, and the reporting of the statistical analyses should be strengthened. In addition, the discussion occasionally overstates the practical implications of relatively small associations observed in a cross-sectional design.

 

I would encourage the authors to address the points below carefully, as the study has the potential to make a useful contribution to the literature.

 

Abstract

• The conclusion that psychosomatic symptoms are “more closely associated” with endocrine status than perceived stress or mood should be interpreted more cautiously. The observed effect sizes are relatively small and the cross-sectional design does not allow causal inferences.
• Consider reporting effect sizes in addition to p-values for the main findings.
• The practical implications could be stated more clearly in the final sentence.

 

 

Introduction

• The Introduction is generally comprehensive but somewhat lengthy. Several paragraphs could be condensed to improve readability.
• The rationale for focusing on psychosomatic symptoms rather than other commonly used athlete-monitoring tools (e.g., wellness questionnaires, recovery scales, training load metrics) could be better justified.
• The authors repeatedly refer to endocrine imbalance and maladaptation. However, the present study does not directly assess overtraining, RED-S, performance decrements, or clinical endocrine dysfunction. The terminology should therefore be used with greater caution.
• A clearer and more explicit statement of the study hypotheses would strengthen the end of the Introduction.

 

 

Methods

• The age range is very broad (14–35 years), and substantial physiological maturation occurs during adolescence. Although age was included as a covariate, further justification is needed regarding whether age adjustment alone is sufficient to account for developmental differences.

 

• Female hormonal status is likely influenced by menstrual cycle phase. While estradiol values were normalized relative to cycle-specific reference medians, it remains unclear how accurately cycle phase was determined and whether potential classification errors could influence the findings.

 

• Given that salivary cortisol was one of the primary endocrine markers investigated, additional information regarding pre-sampling controls (recent exercise, caffeine intake, hydration status and nutritional intake) would be helpful, as these factors may contribute to cortisol variability. Although sampling time was standardized, compliance with awakening-time instructions was not objectively verified.

 

• The justification for categorizing continuous psychological variables into groups (PSS, POMS and HBSC categories) should be clarified. Categorization may reduce statistical power and obscure underlying relationships.

 

• The authors report non-normal distributions and subsequently use GLM analyses. Please clarify whether model assumptions (normality of residuals, homoscedasticity) were evaluated before applying the general linear models.

 

 

 

Results

• Tables 5–7 suggest clear age-related trends in psychological variables. It would be useful to report effect sizes alongside significance values.

 

• The apparent discrepancy between the non-parametric subgroup comparisons (Table 8) and the significant GLM findings (Table 9) should be explained more clearly.

 

• Figures 1 and 2 are relatively simplistic and provide limited information. Consider presenting confidence intervals, individual data points, or adjusted estimated marginal means.

 

• Partial η² values indicate mostly small effects. This should be emphasized when presenting the findings.

 

 

Discussion

• The Discussion is generally well written and appropriately linked to previous literature.

 

• However, some interpretations appear stronger than warranted by the observed associations. The findings indicate statistically significant relationships, but not necessarily clinically meaningful associations.

 

• Alternative explanations should be considered. For example, psychosomatic symptoms and hormonal alterations may both be influenced by unmeasured factors such as training load, recovery status, sleep quality, nutritional status, or competition schedule.

 

• The authors may also wish to consider discussing the known variability of cortisol measurements and the influence of methodological factors on endocrine assessment. Previous work has demonstrated substantial inter-individual variability in cortisol responses and highlighted the importance of controlling factors such as circadian timing and pre-sampling conditions when interpreting hormonal data. See example: https://doi.org/10.1123/ijspp.2018-1004

 

 

Limitations

 

• The authors appropriately acknowledge several limitations; however, I recommend explicitly adding:
• The cross-sectional design.
• Lack of objective training-load data.
• Lack of sleep and recovery measurements.
• Potential influence of menstrual-cycle classification errors.
• Small effect sizes despite statistical significance.

Minor comments

• Several sections would benefit from language editing.
• Some sentences are repetitive, particularly in the Discussion.
• Consider shortening the manuscript slightly by reducing redundancy between Results and Discussion.

Comments for author File: Comments.pdf

Author Response

General comment: The manuscript investigates the associations between endocrine status (salivary cortisol, testosterone and estradiol) and psychological/psychosomatic indicators in elite handball players. The topic is relevant for athlete monitoring and the prevention of maladaptation to training, and the study benefits from a large sample of elite athletes across a wide age range.

Overall, the manuscript is generally well structured and addresses an important applied question. The sample size is a major strength and the inclusion of both endocrine and psychosomatic variables is potentially valuable. However, several methodological and statistical issues require clarification. In particular, the rationale for some analytical decisions, the interpretation of the findings, and the reporting of the statistical analyses should be strengthened. In addition, the discussion occasionally overstates the practical implications of relatively small associations observed in a cross-sectional design.

I would encourage the authors to address the points below carefully, as the study has the potential to make a useful contribution to the literature.

Response: The Authors would like to express sincere thanks to the Reviewer for careful reading and comments, suggestions for improvement in the paper.

 

Comment 1: Abstract - The conclusion that psychosomatic symptoms are “more closely associated” with endocrine status than perceived stress or mood should be interpreted more cautiously. The observed effect sizes are relatively small and the cross-sectional design does not allow causal inferences. Consider reporting effect sizes in addition to p-values for the main findings. The practical implications could be stated more clearly in the final sentence.

Response: The Authors thank these comments of the Reviewer. The effect sizes were added to p-values in the Results section as follows:

“Age-related differences in hormonal levels were assessed using the Kruskal-Wallis test. No significant differences were observed in testosterone levels across age groups in males (p=0.261, η²=0.004 – negligible size effect). In contrast, cortisol levels in males (p=0.042, η²=0.021 – small size effect) and all hormonal parameters in females (testosterone p=0.003, η²=0.041 – small size effect; cortisol p<0.001, η²=0.061 – large size effect; estradiol p<0.001, η²=0.141 – large size effect) differed significantly across age groups, showing increasing trends with age.

Stress, mood, and psychosomatic indicators were also non‑normally distributed (Tables 5–7; Shapiro-Wilk test: males – Perceived Stress Scale p=0.014, Profile of Mood States p<0.001, Health Behavior in School‑aged Children Symptom Checklist p<0.001; females – Perceived Stress Scale p=0.035, Profile of Mood States p<0.001, Health Behavior in School‑aged Children Symptom Checklist p<0.001) and were therefore analyzed using the Kruskal-Wallis test. All indicators showed significant age‑related differences, with decreasing trends observed in both males and females (all p<0.001 – males: Perceived Stress Scale - η²=0.020 – small size effect; Profile of Mood States - η²=0.033 – small size effect; Health Behavior in School‑aged Children Symptom Check-list - η²=0.014 – small size effect, females: Perceived Stress Scale - η²=0.034 – small size effect; Profile of Mood States - η²=0.101 – medium size effect; Health Behavior in School‑aged Children Symptom Checklist - η²=0.081 – medium size effect).

Comparisons of hormonal parameters across stress, mood, and psychosomatic health status categories (Table 8) revealed no statistically significant differences in either sex (Kruskal-Wallis test; perceived stress categories – testosterone: males p=0.338, η²=0.002 - negligible size effect, females p=0.196 - η²=0.001 – negligible size effect; estradiol: females p=0.439 - η²=0.001 – negligible size effect; cortisol: males p=0.501 - η²=0.001 – negligible size effect, females p=0.277 - η²=0.002 – negligible size effect; mood state categories – testosterone: males p=0.338, η²=0.002 - negligible size effect, females p=0.123 - η²=0.010 – small size effect; estradiol: females p=0.691 - η²=0.010 – small size effect; cortisol: males p=0.860 - η²=0.001 – negligible size effect, females p=0.060 - η²=0.022 – medium size effect; psychosomatic health status categories - testosterone: males p=0.668, η²=0.001 - negligible size effect, females p=0.160 - η²=0.006 – negligible size effect; estradiol: females p=0.681 - η²=0.001 – negligible size effect; cortisol: males p=0.476 - η²=0.004 – negligible size effect, females p=0.192 - η²=0.005 – negligible size effect).”

The Abstract was completed with the following sentence about effect sizes:

“However, these associations were characterized by relatively small effect sizes, indicating that psychosomatic symptoms explain only a limited proportion of the variance in hormonal parameters.”

The Abstract was completed with the following final sentence:

“In practical terms, routine monitoring of psychosomatic symptoms alongside hormonal measures may help practitioners to identify early signs of physiological strain and support timely adjustments in training load and recovery strategies.”

 

Comment 2: Introduction - The Introduction is generally comprehensive but somewhat lengthy. Several paragraphs could be condensed to improve readability.

The rationale for focusing on psychosomatic symptoms rather than other commonly used athlete-monitoring tools (e.g., wellness questionnaires, recovery scales, training load metrics) could be better justified.

The authors repeatedly refer to endocrine imbalance and maladaptation. However, the present study does not directly assess overtraining, RED-S, performance decrements, or clinical endocrine dysfunction. The terminology should therefore be used with greater caution.

A clearer and more explicit statement of the study hypotheses would strengthen the end of the Introduction.

Response: The Authors thank this comment, as well. The Introduction section was revised to reduce repetition, clarify the rationale for focusing on psychosomatic symptoms, apply more cautious terminology regarding endocrine responses, and provide a clearer statement of the study hypothesis.

 

Comment 3: Methods - The age range is very broad (14–35 years), and substantial physiological maturation occurs during adolescence. Although age was included as a covariate, further justification is needed regarding whether age adjustment alone is sufficient to account for developmental differences.

Response: The Authors are thankful for this comment of the Reviewer. Although age was included as a covariate to account for developmental differences, and we further examined age‑related patterns using age‑group analyses, although residual maturation effects during adolescence cannot be entirely excluded. The following sentence was added to the Methods section:

“In addition to including age as a covariate in the GLM analyses, age‑group comparisons were performed to explore developmental differences across the studied population.”

 

Comment 4: Female hormonal status is likely influenced by menstrual cycle phase. While estradiol values were normalized relative to cycle-specific reference medians, it remains unclear how accurately cycle phase was determined and whether potential classification errors could influence the findings.

Response: The Authors thank this comment, as well. The Methods section on phase of the cycle was verification was completed as follows:

“The determination of menstrual cycle phase was based on self-reported data obtained through personal interviews with the athletes, including information on cycle length, bleeding patterns, regularity and the date of the last menstruation.”

 

Comment 5: Given that salivary cortisol was one of the primary endocrine markers investigated, additional information regarding pre-sampling controls (recent exercise, caffeine intake, hydration status and nutritional intake) would be helpful, as these factors may contribute to cortisol variability. Although sampling time was standardized, compliance with awakening-time instructions was not objectively verified.

Response: We thank the Reviewer for this comment. Pre‑sampling instructions (including abstaining from eating, drinking and oral activities for at least 30 minutes) were standardized, and samples were collected under resting morning conditions; however, we acknowledge that additional factors such as recent exercise, caffeine intake, hydration and nutritional status, as well as variability in awakening time, may have contributed to cortisol variability, and this limitation has now been more clearly stated in the manuscript. An additional sentence was added to the Limitations as follows:

“In addition, variability in salivary cortisol levels may have been influenced by uncontrolled pre‑sampling factors such as recent physical activity, caffeine intake, hydration status, nutritional intake and variability in awakening time.”

 

Comment 6: The justification for categorizing continuous psychological variables into groups (PSS, POMS and HBSC categories) should be clarified. Categorization may reduce statistical power and obscure underlying relationships.

Response: The Authors thank this comment of the Reviewer. The categorization of psychological variables was applied to facilitate clinically interpretable subgroup comparisons and to align with established cut‑off–based classifications of the instruments, while the potential loss of statistical power and information has now been acknowledged as a limitation in the manuscript as follows:

“The categorization of psychological variables was based on established cut‑offs to facilitate interpretation, although this approach may reduce statistical power and obscure continuous relationships.”

 

Comment 7: The authors report non-normal distributions and subsequently use GLM analyses. Please clarify whether model assumptions (normality of residuals, homoscedasticity) were evaluated before applying the general linear models.

Response: We thank for this comment of the Reviewer. Although the original variables showed non‑normal distributions, assumptions of the general linear models (including normality of residuals and homogeneity of variance) were evaluated and were deemed acceptable, and this is clarified in the manuscript as follows:

“Model assumptions for the general linear models, including normality of residuals and homogeneity of variance, were evaluated using residual diagnostics (including visual inspection of residual plots), and were considered acceptable.”

 

Comment 8: Results - Tables 5–7 suggest clear age-related trends in psychological variables. It would be useful to report effect sizes alongside significance values.

Response: The Authors thank the Reviewer for this helpful suggestion. Effect sizes (η²) have now been reported alongside significance values for age-related differences in both hormonal and psychological variables. The Results section was revised accordingly, and effect sizes were interpreted based on conventional thresholds, indicating predominantly small to moderate effects, with some larger effects observed for specific variables (e.g., cortisol and mood disturbance in females).

“Age-related differences in hormonal levels were assessed using the Kruskal-Wallis test. No significant differences were observed in testosterone levels across age groups in males (p=0.261, η²=0.004 – negligible size effect). In contrast, cortisol levels in males (p=0.042, η²=0.021 – small size effect) and all hormonal parameters in females (testosterone p=0.003, η²=0.041 – small size effect; cortisol p<0.001, η²=0.061 – large size effect; estradiol p<0.001, η²=0.141 – large size effect) differed significantly across age groups, showing increasing trends with age.

Stress, mood, and psychosomatic indicators were also non‑normally distributed (Tables 5–7; Shapiro-Wilk test: males – Perceived Stress Scale p=0.014, Profile of Mood States p<0.001, Health Behavior in School‑aged Children Symptom Checklist p<0.001; females – Perceived Stress Scale p=0.035, Profile of Mood States p<0.001, Health Behavior in School‑aged Children Symptom Checklist p<0.001) and were therefore analyzed using the Kruskal-Wallis test. All indicators showed significant age‑related differences, with decreasing trends observed in both males and females (all p<0.001 – males: Perceived Stress Scale - η²=0.020 – small size effect; Profile of Mood States - η²=0.033 – small size effect; Health Behavior in School‑aged Children Symptom Check-list - η²=0.014 – small size effect, females: Perceived Stress Scale - η²=0.034 – small size effect; Profile of Mood States - η²=0.101 – medium size effect; Health Behavior in School‑aged Children Symptom Checklist - η²=0.081 – medium size effect).”

 

Comment 9: The apparent discrepancy between the non-parametric subgroup comparisons (Table 8) and the significant GLM findings (Table 9) should be explained more clearly.

Response: The Authors thank this insightful comment, too. The discrepancy reflects methodological differences, as non‑parametric subgroup comparisons assess unadjusted group differences, whereas GLM analyses account for covariates (e.g., age) and allow evaluation of effects within a multivariable framework, which may reveal associations not apparent in unadjusted analyses; this has now been clarified in the manuscript. The Results section was completed with the following:

“The differences between non‑parametric subgroup comparisons and GLM results reflect the adjustment for covariates and the multivariable modeling approach in GLM analyses, which may detect associations not evident in unadjusted comparisons.”

 

Comment 10: Figures 1 and 2 are relatively simplistic and provide limited information. Consider presenting confidence intervals, individual data points, or adjusted estimated marginal means.

Response: We thank the Reviewer’s comment and suggestion, Figures 1 and 2 were corrected as follows:

 

“Figure 1. Figure 1. Relative salivary cortisol levels (%) across psychosomatic symptom severity (HBSC‑SCL categories) in male handball players - data are presented as estimated marginal means (±95% confidence intervals) and derived from general linear models adjusted for age.”

 

 

“Figure 2. Salivary testosterone levels (pg/ml) across psychosomatic symptom severity (HBSC-SCL categories) in female handball players - data are presented as estimated marginal means (±95% confidence intervals) and derived from general linear models adjusted for age.”

 

Comment 11: Partial η² values indicate mostly small effects. This should be emphasized when presenting the findings.

Response: The Authors thank the Reviewer for this comment. We have revised the manuscript to more clearly emphasize that the observed associations were characterized by predominantly small effect sizes and have interpreted the findings more cautiously in both the Results and Discussion sections.

In the Discussion section:

“In male athletes, higher levels of psychosomatic symptoms were associated with elevated cortisol levels, while in female athletes, associations were observed with testosterone levels (these results should be interpreted with caution, as cortisol levels may have been influenced by variability in sampling conditions, particularly the lack of strict control over wake‑up time). However, these associations were characterized by relatively small effect sizes, indicating limited explanatory power.”

“In summary, the findings of this study indicate that psychosomatic symptom severity is more closely associated showed statistically significant but modest associations with endocrine status than perceived stress or mood disturbance in elite handball players.”

 

Comment 12: Discussion - The Discussion is generally well written and appropriately linked to previous literature. However, some interpretations appear stronger than warranted by the observed associations. The findings indicate statistically significant relationships, but not necessarily clinically meaningful associations.

Alternative explanations should be considered. For example, psychosomatic symptoms and hormonal alterations may both be influenced by unmeasured factors such as training load, recovery status, sleep quality, nutritional status, or competition schedule.

Response: We thank the Reviewer for this comment. The interpretation of the findings has been revised to emphasize the modest magnitude and limited clinical relevance of the observed associations, and the Discussion has been expanded to consider alternative explanations, including the potential influence of unmeasured factors such as training load, recovery status, sleep quality, nutritional status, and competition schedule. The Discussion section was completed as follows:

“The most important finding of the present study is that psychosomatic symptom severity, assessed using the HBSC checklist, demonstrated the most consistent, albeit modest, association with hormonal parameters. In male athletes, higher levels of psychosomatic symptoms were associated with elevated cortisol levels, while in female athletes, associations were observed with testosterone levels (these results should be interpreted with caution, as cortisol levels may have been influenced by variability in sampling conditions, particularly the lack of strict control over wake‑up time). However, these associations were characterized by relatively small effect sizes, indicating limited explanatory power and suggesting limited practical or clinical relevance. In addition, the observed associations may be influenced by unmeasured factors such as training load, recovery status, sleep quality, nutritional status, or competition schedule, which were not assessed in the present study and may contribute to both hormonal variation and psychosomatic symptom reporting.”

 

Comment 13: The authors may also wish to consider discussing the known variability of cortisol measurements and the influence of methodological factors on endocrine assessment. Previous work has demonstrated substantial inter-individual variability in cortisol responses and highlighted the importance of controlling factors such as circadian timing and pre-sampling conditions when interpreting hormonal data. See example: https://doi.org/10.1123/ijspp.2018-1004

Response: We thank for valuable suggestion of the Reviewer. The Discussion has been expanded to acknowledge the well‑documented variability of cortisol measurements and the influence of methodological factors, including circadian timing and pre‑sampling conditions, on endocrine assessment, with reference to the recommended literature. The Discussion section was completed as follows:

“The most important finding of the present study is that psychosomatic symptom severity, assessed using the HBSC checklist, demonstrated the most consistent, albeit modest, association with hormonal parameters. In male athletes, higher levels of psychosomatic symptoms were associated with elevated cortisol levels, while in female athletes, associations were observed with testosterone levels (these results should be interpreted with caution, as cortisol levels may have been influenced by variability in sampling conditions, particularly the lack of strict control over wake‑up time). In addition, salivary cortisol levels are known to exhibit substantial inter‑individual variability and are highly sensitive to methodological factors, including circadian timing, recent physical activity, and pre‑sampling conditions. Previous research has high-lighted the importance of controlling these factors when interpreting cortisol measurements in athletic populations [26].”

 

Comment 13: Limitations - The authors appropriately acknowledge several limitations; however, I recommend explicitly adding:

The cross-sectional design.

Lack of objective training-load data.

Lack of sleep and recovery measurements.

Potential influence of menstrual-cycle classification errors.

Small effect sizes despite statistical significance.

Response: The Authors thank the suggestions for the Limitations section. This is the revised Limitations section:

“The present study has several limitations. The cross‑sectional design precludes conclusions regarding causal relationships between psychological indicators and hormonal parameters. Although saliva sampling provides a practical and non‑invasive method for assessing endocrine status, variability related to sampling conditions - particularly the lack of strict control over wake‑up time, which is a key determinant of cortisol levels due to its pronounced circadian rhythm - may have affected the measurements. In addition, variability in salivary cortisol levels may have been influenced by uncontrolled pre‑sampling factors such as recent physical activity, caffeine intake, hydration status, nutritional intake and variability in awakening time. Furthermore, objective measures of training load, sleep quality, and recovery status were not available, although these factors may influence both hormonal responses and psychosomatic symptoms. In female athletes, classification of menstrual cycle phase was based on self‑reported data, which may have introduced some degree of misclassification error.

Although several associations reached statistical significance, the observed effect sizes were relatively small, indicating limited explanatory power and practical relevance.

Psychological variables were assessed using self‑report questionnaires, which are inherently subject to reporting bias and may not fully capture underlying psychological states. The categorization of psychological variables was based on established cut‑offs to facilitate interpretation, although this approach may reduce statistical power and obscure continuous relationships.

The findings reflect relative hormonal variations within a specific population of elite athletes and should therefore not be interpreted as indicators of clinical risk or endocrine dysfunction.”

 

Comment 13: Minor comments - Several sections would benefit from language editing.

Some sentences are repetitive, particularly in the Discussion. Consider shortening the manuscript slightly by reducing redundancy between Results and Discussion.

Response: We thank this helpful suggestion, too. The manuscript has been thoroughly revised for language and clarity, and repetitive phrasing - particularly in the Discussion - has been reduced to improve readability and overall flow.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

No further comments. Thank you for your work.