Search strategies

To provide a wide, complete and thorough systematic review over hormonal aspects in OTS/NFOR/FOR, PRISMA protocol for systematic reviews was used for the study design and a systematic search was conducted through the electronic PUBMED, MEDLINE (Ebsco), and COCHRANE databases, from their earliest records to November 8th, 2017. Searched expressions were: (1)overtraining, (2)overreaching, (3)overtrained, (4)overreached or (5)underperformance, and (plus) (a) hormone, (b) hormonal, (c) endocrine, (d) adrenal, (e) cortisol, (f) GH, (g) ACTH, (h) testosterone, (i) IGF-1, (j) TSH, (k) T4, (l) T3, (m) LH, (n) FSH, (o) prolactin and (p) IGFBP-3, in a total of 85 expressions searched.

We also analyzed articles that were mentioned within identified studies whenever OTS/NFOR/FOR or a similar disorder was described (e.g. cortisol and adrenocorticotropic hormone (ACTH) profiles and over-trained athletes).

Data extraction

All studies were evaluated after removal of duplicate articles, according to: 1) authorship, 2) publication date, 3) studied population, 4) definition of overtraining and overreaching syndromes, 6) study design and methods, 7) methods of hormone assessment, 8) results, 9), conclusions and 10) study variables and bias.

Eligibility criteria

Inclusion criteria were: 1) whole article written in English; 2) hormone profile (basal, stimulated, or both) and correlation to OTS/NFOR/FOR states as an outcome (but not necessarily the primary outcome); 3) absence of influence of any hormone therapy; 4) absence of confounding diseases that would lead to an impaired hormonal status caused by the disorder itself, which would also exclude the diagnoses of OTS/NFOR/FOR; 5) population with proper diagnosis of OTS/NFOR/FOR; and 6) studies with humans.

Quality assessment

The abstract of each of researched study was analyzed by one of the authors (F.A.C.) and assessed for inclusion according to the eligibility criteria. The studies that fulfilled the inclusion criteria were evaluated with regard to their rationale, study design, primary outcomes, assessment of OTS/NFOR/FOR diagnosis, statistical analysis, results, discussion and conclusions. Studies that presented any bias in methodology, results, or interpretation of the exposed data, that could interfere in the analysis were excluded. Studies whose methodology could not be assessed by the abstract had the full text analyzed, prior to the decision of the acceptance for the systematic review.

Statistical analysis

Due to the heterogeneity of the studies that aim to correlate hormone responses and OTS/NFOR/FOR, a meta-analysis, our initial idea, was unfeasible. Moreover, since several variables were taken into account, including the type of sports performed, type of tests and stimulus performed, and whether hormones were analyzed basally, and during a resting period or resting after an intensive training load, analysis of each aspect was performed separately, and in several different combinations, including studies that performed equivalent methods, although not similar, including all tests stimulated by exercises, regardless of the type of stimulation, tests performed in endurance sports, regardless of the type of sport, and resting combined with basal hormones, regardless of the previous exercise stimulation. Data was also analyzed according to the number of studies investigating each type of sport, the number of athletes with OTS/NFOR/FOR and healthy athletes, and the number of studies using each type of hormone stimulation test were quantified. Study results were analyzed in terms of percentage of type of responses for each of the tests performed. Once meta-analysis was not performed, t-tests, one-way ANOVA and non-parametric tests were not employed, and therefore a statistical analysis program was not necessary.

Selection of the studies

The initial search yielded 835 articles. The sum of the searches for “overtraining”, “overreaching”, “overtrained”, “overreached” and “underperformance”: 1. with “catecholamines” yielded 62 articles; 2. With “ACTH” yielded 29 articles; 3. with “cortisol” yielded 161 articles; 4. with “adrenal” yielded 60 articles; 5. “GH” yielded 13 results; 6. with “IGF-1” yielded 19 articles; 7. with “IGFBP-3” yielded four articles; 8. with “testosterone” yielded 100 articles; 9. with “LH” yielded nine articles; 10. with “FSH” yielded four articles; 11. with “TSH” yielded four articles; 12. with “T4” yielded 12 articles; 13. with “T3” yielded 13 articles; 14. with “prolactin” yielded 10 articles; 15.with “hormone” yielded 195 articles; 16. with “hormonal” yielded 92 articles; and 17. with “endocrine” yielded 49 articles.

Among the yielded articles, the reasons of the exclusions were due to: (1) Review of FOR/NFOR/OTS – 57 articles; (2) Performed in animals – 35 articles; (3) Nutritional supplementation for FOR/NFOR/OTS – 20 articles; (4) Hormonal physiological adaption to exercises – 13 articles; (5) Other markers of FOR/NFOR/OTS than not hormones – 14 studies; (6) Markers of stress during training – 09 articles; (7) Other dysfunctions than FOR/NFOR/OTS evaluated (depression, amenorrhea) – 06 studies; (8) Nutritional management of FOR/NFOR/OTS – 06 studies; (9) Improper criteria for diagnosis of FOR/NFOR/OTS – 05 articles; (10) Stimulation therapies effects on FOR/NFOR/OTS (electrical, cryotherapies) – 04 articles; (11) Post-hoc analysis – 03 articles; (12) Non-FOR/NFOR/OTS review – 03 articles; (13) Training management for FOR/NFOR/OTS – 02 articles; (14) Predictive markers of FOR/NFOR/OTS – 02 articles; (15) Pre-stressful events hormonal alterations – 02 articles; (16) Markers of competition performance (but not for FOR/NFOR/OTS) – 01 article; (17) Non-hormonal adaptions to exercise – 01 article; (18) Non-sport FOR/NFOR/OTS – 01 article; (19) In-vitro study – 01 article; and (20) Guideline – 01 article. The total of excluded articles that were not due to duplication were 179 studies, which means among the non-duplicate yielded articles, 82.5% were excluded, whereas 38 (17.5%) were included.

Study characteristics

Type of performed tests

Among the 38 selected studies, 24 (63.2%) evaluated resting hormone levels after an induced NFOR/FOR state, 17 (44.7%) tested basal hormone levels and 16 (42.1%) evaluated acute hormone responses to stimulation tests. Among these, nine (23.7%) evaluated both basal and stimulated hormones, whereas seven (18.4%) tested both acute hormone responses to functional tests and resting levels after an overload training program.

Induction of NFOR/FOR state in healthy athletes was performed either by an OTP (22 studies) or after a Competition Season Period (CSP) (three studies), followed by a proper diagnosis of OTS/NFOR/FOR (reduction of at least 10% of the previous performance). One study [3] performed both OTP and CSP.

Conversely, among the stimulation tests performed in previously affected athletes, 11 (68.8% of all stimulation tests) were carried out with maximal exercise (ME), three (18.8%) with two-bout protocol exercises (TBE) and two (12.5%) with an insulin tolerance test (ITT). Seven of the studies performed two types of stimulation tests.

Five studies [10, 28, 30, 35, 36] performed both resting levels after an OTP and hormones responses to ME in NFOR/FOR induced athletes by OTP, one [19] performed both basal and stimulated hormones by ITT, and one [18] performed ME response after an OTP, but not basal levels; the seven studies that explored acute responses after an OTP means 33.3% of total of studies that performed OTP; the six studies that performed ME after an OTP represents 54.6% of total studies that employed ME, as the other studies did ME in previously affected athletes without any previous type of overload training. Figure 1 summarizes the types of athletes and tests performed in the selected studies.

Regarding the object of comparison for analysis purposes, 20 studies (52.6%) compared resting hormone levels of FOR-induced athletes by OTP/CSP with previous resting levels; 18 (47.4%) compared basal levels of affected athletes with basal levels of healthy athletes; 12 (31.6%) compared acute hormone responses to stimulation tests between affected and healthy athletes; 11 (28.9%) compared compared resting hormone levels of FOR-induced athletes with resting levels of healthy athletes; and four (10.5%) compared acute hormone response to stimulation tests with basal levels of the same affected athletes. From the 38 studies, 21 (55.2%) performed more than one type of comparison (eg. Analysis between basal levels of the same affected athletes and also between affected and healthy athletes).

Hormone findings from the selected studies

Systematic search

The aim of this systematic review is to evaluate hormonal aspects of OTS/NFOR/FOR already published. Therefore, after the search for a wide number of expressions and hormones, only 38 met the criteria, as many expressions yielded the same studies. Moreover, 82.5% of the studies were descriptive, did not perform hormone tests, or therefore did not add new information or evidence to correlation between OTS/NFOR/FOR and hormone profile. The systematic search was expanded beyond the PRISMA guidelines for systematic reviews in order to achieve all the studies that correlate hormone levels and OTS/NFOR/FOR, due to the wide variety of expressions and study designs.

As stated in the results section, it was unfeasible to analyze each sport separately due to the small number of subjects of each study, wide distribution of sports performed, heterogeneity regarding most aspects of the studies, such as inclusion criteria, and precise definition of who were affected by OTS/NFOR/FOR.

Interestingly, more reviews on hormonal aspects of OTS/NFOR/FOR were found (55 reviews) than the number of selected studies (44.7% more than original articles). Moreover, five studies improperly diagnosed OTS/NFOR/FOR based on hormones alterations, such as reductions of 30% or more of T/C ratio, regardless of the performance status.

Types of parameters and tests performed

There were three types of assessed hormones: 1. basal hormones in previously affected athletes; 2. resting levels after an overload training; and 3. acute hormone responses to stimulation tests in OTS/NFOR/FOR individuals.

Among the 38 selected studies, 26 different hormones and hormone-related parameters were assessed, both basally and stimulated. Interestingly, the stimulation tests performed were not those recommended by endocrine societies and have not been validated previously, except two, that performed ITT; indeed, the oldest study found by this systematic review, performed by Barron et al. [38] and published in 1985 performed ITT and found blunted ACTH and cortisol levels in OTS/NFOR athletes; however, no further studies that tried to reproduce these findings were found. ITT is a gold-standard functional test recommended by Endocrine societies to evaluate stress hormone response to a simulation of a stressful situation, the hypoglycemia. This test requires the whole axis integrity in order to provide normal responses; and, regardless of the affected level of each axis, the lack of response mimic the real-life stressful situations, which may lead to decreased performance in extreme sports sessions, when integrity of stress-related axis are required.

Although not supported by Endocrine societies, some authors showed that hormonal responses to intense exercises (ME) seemed to be more appropriate to evaluating OTS/NFOR/FOR, which is confirmed by the findings of this systematic review. Thus five different kinds of stimulation tests were used for the evaluation of hormonal responsiveness, including three that evaluated acute response tests (ME, TBE and ITT), and two that measured in resting after a training overload (OTP and CSP). According to the most recent OTS/NFOR/FOR guidelines [1], the most cited test among the reviewed studies was TBE; however, we observed that ME, and not TBE, was the most studied functional test, and was able to show significant differences between OTS/NFOR/FOR and healthy athletes. However, none of the studies established a proposed cutoff for any of the tested hormones, and most of the conclusions recommend further studies to establish validated cutoff and criteria.

The heterogeneity of ME protocols, the lack of cutoff for the tests hormones (despite of the altered hormone findings), and the small number of participants in most studies (as it is not easy to recruit athletes affected with OTS/NFOR/FOR in the current moment of the disease) lead to the fact that is is possibl but uncertain whether the conducted tests are useful for the accurate diagnosis of OTS/NFOR/FOR. Further studies should explore the findings of this systematic review and perform validated tests by Endocrine societies, such as ITT, and standardized tests by the latest guideline [1], such as TBE, and provide correlation between responses of thes two types of tests, as well as attempt to establish reliable cutoffs for each hormone in order to provide a new biomarker for OTS/NFOR/FOR.

Types of sports practiced

OTS/NFOR/FOR related to endurance sports was analyzed by 35 (92.1%) studies, whereas resistive exercises were only analyzed by three (7.9%), accordingly with the lack of studies described by the latest guidelines [1].

Regarding endurance exercises, despite the potentially high prevalence of OTS/NFOR/FOR among triathlon athletes, only six studies performed tests in this population, whereas other sports which OTS/NFOR/FOR has been less described have been perhaps disproportionately studies.

Unlike in endurance sports, that showed reduction in catecholamines levels, an increase was observed in weight lifters affected by OTS/NFOR, as seen in the three studies that were performed in resistive athletes (weight-lifters), which distinct results could be observed from the other studies, once the two studies that evaluated catecholamines found increased levels, and decreased beta-receptor concentrations (down-regulation) in muscles were found in one, with consequent decreased sensitivity to catecholamines. This finding is corroborated by Fry et al. [17], whose correlation coefficients suggested decreased responsivity of skeletal muscle to sympathetic nervous system activity and therefore corroborates the sympathetic aspect of the resistive FOR. Therefore, the named sympathetic OTS/NFOR/FOR for resistive athletes [1, 39] may be appropriate, although further studies are necessary to corroborate these findings. as there is a lack of previous controlled studies with resistive exercises.

Hormones and OTS/NFOR/FOR

The role of hormones in overtraining and overreaching syndromes

The lack of alterations in basal and resting levels in OTS/NFOR/FOR athletes may be explained by the high adaptive capacity of athletes to extreme environments, which leads to shorter periods of recovery after physical exertion. The basal hormone release may not be altered in these subjects, distinctly from the possible impaired capacity to respond to stressful situations.

Indeed, despite of the limitations found in this systematic review, the consistent findings of blunted ACTH, GH and prolactin levels observed in all stimulation tests deserve attention. Hormones and glands physiology present similarities among them in terms of basal and stimulated hormone production and release. In this case, initial stages of glands dysfunctions tend to preserve basal hormone production, but present impaired responses to stimulations, such as stressful situations. Hormonal alterations may be therefore only seen in acute responses to functional tests, but not in basal and in resting levels, as observed in relative adrenal insufficiency, GH deficiency and pre-diabetes and initial diabetes. Relative adrenal insufficiency shows as a lack of enough raise in cortisol levels to ITT or to ACTH-stimulation test; GH deficiency is diagnosed when a blunted GH is observed to ITT; and pre-diabetes or initial diabetes are more accurately diagnosed by Oral Glucose Tolerance Test, which provides earlier information about disglycemia than fasting serum glucose, once relative hypoinsulinemia with consequent hyperglycemia in response to oral glucose is observed. In fact, severe adrenal insufficiency and final stage of diabetes are required to present basal hypocortisolism and fasting hyperglycemia with inappropriate hypoinsulinemia, respectively. TRH stimulation test for thyroid function and LHRH stimulation test for ovary function are able to show depletion of these glands prior to basal levels as well. Therefore, it is possible, if not probable, that compromised glands with relative hormone dysfunction are present in OTS, although further studies are required, particularly linking acute hormonal responses to exercises with functional tests standardized by endocrine societies.

Notably, hormonal changes in OTS/NFOR/FOR are probably not triggers of these disorders, but may play a role in the worsening and perpetuation of OTS/NFOR/FOR symptoms. Thus once identified as markers, hormones should not be replaced; instead, their recovered levels could be used as markers of improvement in OTS/NFOR/FOR. Additionally, once basal levels are altered, a diagnosis of OTS/NFOR/FOR is unlikely, as these conditions should not be diagnosed in the presence of endocrine alterations [1, 4, 28].

Although acute responses to stress in OTS/NFOR/FOR athletes and resting levels responses to NFOR/FOR state were also analyzed together, they may present distinct results, once they were measured in different situations, allied to the fact that OTS may differ biochemically from NFOR/FOR states. Indeed, one study showed different results between OTS and NFOR/FOR [9], once in OTS, GH and ACTH failed to raise in response to a TBE, whereas in NFOR/FOR athletes both hormones markedly raised in response to TBE. It was also observed that studies with previously affected athletes tend to show more pronounced altered responses than studies that induces FOR/NFOR, perhaps due to the above-mentioned possible difference between FOR (when overload is performed, a FOR/NFOR state is induced, not OTS) and OTS (most studies with previously affected subjects tended to include OTS, not FOR/NFOR athletes).

Differential diagnosis

Many dysfunctions that may confound the proper OTS/NFOR/FOR diagnoses have to be excluded in athletes that also lead to worsened performance and fatigue, including: 1. Sleep disorders; 2. Site-specific or sub-clinical diffuse inflammation (non-alcoholic steatohepatitis – NASH, metabolic syndrome, dyslipidemia); 3. Frank hormonal dysfunctions; 4. Autoimmune diseases (rheumatoid arthritis, systemic lupus erythematosus, ankylosing spondylitis); 5. Vitamins deficiency (vitamin B12, B1, D); 6. Psychiatric conditions (bipolar disorder, depression, anxiety disorders); 7. Lung dysfunctions (asthma, idiopathic pulmonary fibrosis, tuberculosis) and 8. Heart conditions (persistence space of double interatrial septum, heart failure) Most of the above listed differential diagnosis are not explicitly exposed in the latest OTS/NFOR/FOR guideline [1].

Fatigue is not always linked to reduced performance, and hormonal differences may be observed between non-reduced performance and reduced-performance fatigued athletes. Aubry et al. [2] compared athletes acutely fatigued without decrement of performance and FOR athletes, and catecholamines were found to be reduced in FOR compared to the non-FOR fatigued subjects.

Hormonal markers of OTS/NFOR/FOR: Specific for these conditions?

Some alterations found in OTS/NFOR/FOR athletes may be resulted from overload training, regardless of the performance state, and therefore be also observed in healthy athletes. Indeed, Uusital et al. [28] performed an Experimental Training Group (ETG) with massively intense exercises (above lactate threshold) compared to control group. Cortisol and noradrenaline were decreased in the ETG in both induced FOR (n = 5) and non-FOR (n = 4), compared to control, which means that altered findings may not always differentiate OTS/NFOR/FOR, but indicate an excess of training. The proposed markers of OTS/NFOR/FOR were also contested by Hoogeveen et al. [30], who showed that hormonal changes were the same between OTS and healthy athletes, concluding that all the alterations were related to physiological adaptions, not to reduced sports performance.

Furthermore, despite of the systematic search for decreased total T and T/C ratio as markers of worsened performance, Hoogeveen et al. [30] showed that decreased total T, increased cortisol and decreased T/C ratio failed to predict reduction in performance.

Limitations

Some remarkable limitations in studies that evaluated hormonal levels in OTS/NFOR/FOR were: 1) The impossibility of performing a meta-analyses owing to the wide variety of methodologies; 2) Wide variety of evaluated sports; 3) Lack of standardization criteria to differentiate between OTS, NFOR and FOR; 4) Lack of enough studies comparing OTS/NFOR/FOR with healthy athletes and healthy sedentary individuals, in order to evaluate physiological hormone adaption via training; 5) Small numbers of participants in most studies; 6) Lack of studies with sympathetic OTS/NFOR/FOR (such as in weight-lifters); and 7) Lack of standardized stimulation tests endorsed by endocrinology societies, to enforce the evidence strength.

Besides the limitations of the selected studies, this systematic review also presented some limitations. First, if the basic PRISMA protocol of search for systematic reviews was followed, only 12 studies would have been selected. Therefore, an expanded search that went beyond an usual systematic review had to be performed. Second, a more comprehensive analysis for more precise conclusions was unable to be performed, due to the unexpected differences in methodology, even when types of sports or tests were similar. And third, we found less consistent data then expected, which did not allow a more complete systematic review in the field.

Final discussion

To date, this is the first systematic review that evaluated hormonal aspects of OTS./NFOR/FOR. The large number of studies that compared basal levels between affected and healthy athletes may lead to the conclusion that basal levels of hormones are definitely not good markers of OTS/NFOR/FOR. Conversely, stimulated hormones, particularly acute responses to stressful conditions may be good predictors of underperformance syndromes, although OTS/NFOR/FOR is a complex disorder, with multi-etiologic pathophysiology, and therefore hormones dysfunctions are unlikely to be the only etiology of OTS/NFOR/FOR.

The overall conflicting findings of the present systematic review reflects the complex relation between hormones and overtraining syndrome. The main issue regarding hormone findings in OTS is the causality relationship and the real underlying reason that leads to decreased performance. Whenever the hormone dysfunction is likely the primary cause for the worsened performance, OTS is excluded, once OTS requires exclusion of endocrine disorders to be diagnosed. Conversely, OTS may lead to dysfunctional hormones, as observed [1–38], and identifying whether these dysfunctions are secondary to OTS is challenging. As endocrinologists, we stated that generally, whenever a frank and severe hormone dysfuncion was found, the diagnosis of OTS is unlikely, once OTS tend to induce mild changes in hormone, not important changes, and since the hormone dysfuncion probably explains all the clinical findings, and organic etiologies for the dysfuncion should be searched. Finally, who defines the diagnosis is the underlying etiology: whenever the hormone dysfuncions do not fully explain the clinical presentation, and seem to be a consequence of the OTS, OTS is the diagnosis; contrariwise, if the clinical presentation is attributable to the hormone dysfunctions, the replacement or suppression of the affected hormone should be approached prior to the diagnosis of OTS.

Finally, it may be too early to claim that GH, ACTH and prolactin responses are undoubtedly blunted in acute responses in affected individuals, once not all studies found the same results, studies performed slightly different protocols for ME and TBE, and small number of subjects were included. For practical purposes, whenever an athlete is suspected for OTS/NFOR/FOR, stimulation tests could be performed in order to find possible relative failure of the adrenals or the pituitary, although only standardized tests by Endocrine societies (ITT) are enough to provide these diagnosis.

Further studies should be performed with athletes from the sports that are mostly described in OTS/NFOR/FOR, such as triathlon, cycling and long distance runners, standardization of ME should be employed, both ME and ITT should be performed and correlated, and a specific control group of healthy athletes that practice the equivalent intensity and volume of training should also perform the exact same tests, to strength the level of the found data.