Introduction

Physical exercise produces beneficial health effects (^{Lubans, Foster, & Biddle, 2008}), improving sense of well-being and state of mind (^{Ekkekakis & Acevedo, 2006}). People who exercise regularly tend to have higher levels of health and less stress, irrespective of the type of exercise (^{Jiménez, Martínez, Miró, & Sánchez, 2008}). Other benefits that can be seen are, relief from anxiety, antidepressant effects, increased self-esteem, self-efficacy and emotional regulation (^{Caracuel & Arbinaga, 2010}). However, in some cases, taking part in regular physical exercise can become a behavioural dependence leading to psychological addiction (^{Adams, Miller, & Kraus, 2003}) and a deterioration of state of mind (^{Guszkowska & Rudnicki, 2012}; ^{Jones & Tenenbaum, 2009}).

Exercise addiction has been identified as having characteristics in common with DSM-5 criteria for behavioural addictions (^{Márquez & De la Vega, 2015}). Using the following criteria: a) Tolerance, an increase in the amount of exercise to achieve the desired effect; b) Withdrawal, negative effects are experienced in the absence of exercise, such as anxiety, irritability or sleep problems; c) Lack of control and failure in attempts to reduce or cease exercise; d) Intention effects, an inability to adjust to a set routine; e) Excessive dedication to preparing for, engaging in, and recovering from exercise; f) Reduction of other social, occupational and / or recreational activities; g) Continuance, continuing to exercise despite knowing that this activity is creating problems.

Researchers have studied the effects of different sociodemographic variables on exercise dependence, but results are not conclusive. For example, studies report differences in exercise dependence in men and women depending on the assessment instrument used (^{Weik & Hale, 2009}): men show high levels of dependence when using the Exercise Dependence Scale (EDS), whereas the Exercise Dependence Questionnaire (EDP) gives an opposite result with women having greater dependence. Some authors even report no gender differences (^{Allegre, Therme, & Griffiths, 2007}; ^{Lindwall & Palmeira, 2009}; ^{Reche, Martínez-Rodríguez, & Ortín, 2015}). Also, there is some controversy about the age for the onset of the disorder. Most researchers agree that exercise dependence has a greater influence on young athletes (^{Allegre et al., 2007}; ^{Lindwall & Palmeira, 2009}; ^{González-Cutre & Sicilia, 2012}). Others, however, have found no associations (^{Klein et al., 2004}).

Effects of different basic sport variables have also been studied. In relation to sport modality (individual or team), ^{Reche et al. (2015}) show similar levels of dependence and ^{Modolo, Mello, Gimenez, Tufik and Antunes (2009}) report that there is a greater dependence in amateur athletes than in professional athletes. ^{Sicilia and González-Cutre (2011}), using the Spanish version of EDS-Revised (EDS-R), found that athletes who exercise more than 3 days a week (more sport dedication) scored higher on all subscales of dependence compared to athletes who took part less frequently. Other studies show that participants who do more than four hours of physical exercise per week show a higher overall average of exercise dependence compared to those who do less (^{Reche & Gómez, 2014}; ^{Reche et al., 2015}). So, exercise adherence and regularity could be considered as presenting characteristics. However, ^{Kjelsas, Augestad and Götestam (2003}) do not consider this enough to justify exercise dependence, and so a resolution to this dysfunctional aspect is still something open to discussion.

On the other hand, evidence has suggested that high levels of exercises may be responsible for the development of overtraining syndrome, a state which can lead to burnout (^{Silva, 1990}). From the American Medical Society for Sports Medicine (^{DiFiori et al., 2014}), “burnout is part of a spectrum of conditions that includes overreaching and overtraining” (p. 5). Meanwhile, ^{Cervelló (2003}) established a correspondence between the process of overtraining and burnout, in a way that establishes a parallel in the appearance process of both syndromes. Stress is the trigger element of both processes, but the author indicates that the interpretation that the athlete makes of this is what differentiates the development of burnout from overtraining syndrome.

The first definitions of burnout in sport emerged in the 80s, but it was not until the late 90s that a consensus was established about what burnout was within the sporting context (^{Cresswell & Eklund, 2006}; ^{Gould, Udry, Tuffey, & Loehr, 1996}). According to ^{Goodger, Gorely, Lavallee and Harwood (2007}), the conceptual unification was achieved after a three-dimensional definition was proposed by ^{Raedeke (1997}), characterised by: emotional /physical exhaustion (EFE), sport devaluation (D) and reduced sense of accomplishment (RSA). EFE is defined as feelings of extreme fatigue (physical and psychological) most of the time. D includes negative attitudes toward sport practice. Finally, RSA is reduced accomplishment in terms of sport performance.

Several studies have examined the relationship of burnout or its dimensions with different demographic and sport variables such as gender, age, type of sport, training volume and years of participation. Recent research has found differences in relation to gender, showing that women have the highest values of burnout (^{Heidari, 2013}; ^{Nafian, Karimi, Ghasemi, & Zahrabi, 2014}) and EFE (^{De Francisco, Garcés de los Fayos, & Arce, 2014}). Regarding age, research has yielded conflicting results. ^{Raedeke and Smith (2001}) reported higher levels of burnout in young athletes, ^{Pedrosa and García-Cueto (2014}) found higher levels of burnout in adult athletes and ^{De Francisco et al. (2014)} was unable to find any relationship between age and the different dimensions of burnout. As for the type of sport, classical authors (^{Coakley, 1992}; ^{Smith, 1986}) stated that burnout seems to be more common in athletes of individual sports compared with those who participated as part of a team. However, ^{Gustafsson, Kenttä, Hassmén, and Lundqvist (2007}) using a sample of adolescents failed to confirm this. Positive relationships between burnout (^{Kellmann & Kallus, 2001}; ^{Kenttä, Hassmén, & Raglin, 2001}) or with its EFE dimension (^{De Francisco et al., 2014}), and load and / or training volume have been found, but not with the amount of years of participation (^{Holden, Keshock, Forester, Pugh, & Heitman, 2016}).

In the light of these results, this study aims to provide data on the prevalence of symptoms of exercise dependence and burnout, and to investigate the relationship between the two variables since it is unknown and may be useful for the development of prevention strategies within the sport context. Finally, the influence of different sociodemografic and sport variables on dependence and burnout will be analysed, as they have generated controversy in previous research.

Method

Results

Table 1 shows the descriptive analysis related to the physical exercise dependence of the participants, where 8.7% show risk of exercise dependence, 61.5% show symptoms and 29.8% are asymptomatic. Regarding gender, the percentages are similar (8.5% male and 8.8% female), however, participants aged under 18 years of age show a higher percentage of exercise dependence (10.8% under the age of 18 and 6.8% over 18). Also, in the Table 1, values of chi-square statistic are included in order to check whether the observed differences reach statistical significance or not.

Table 1: Descriptive data on the prevalence of symptoms of dependence on physical exercise by gender and age group, expressed in percentages (number of subjects in parenthesis). 

In relation to the dimensions of burnout, the percentages for suffering from them are between 4% for EFE and 4.9% for RSA and D. Regarding gender, men show higher rates than women in all dimensions, showing major differences in D (0.8% for women vs. 6.6% for men). Differentiating by age group, the percentages are equal for EFE (4.2%) but for the remaining dimensions, over 18’s show a higher percentage (6.4% and 5.9% for RSA and D vs. 3.3% and 3.8%). In table 2, the rest of the risk level percentages for each dimension (low, moderate and high) are also displayed, as well as the values of the chi-square statistic-

To analyse the relationship between exercise dependence and burnout with the sociocultural sport variables, a comparison of the means was carried out between exercise dependence characteristics and dimensions of burnout by gender of athletes, age, sport dedication and type of sport.

Table 2: Descriptive data of the prevalence of the dimensions of burnout in athletes by gender and age group, expressed in percentage (number of subjects in parenthesis). 

The results show statistically significant differences in exercise dependence in terms of age (t = 2.66, p = .001), showing the youngest (under 18 years of age) having an increased risk of exercise dependence. There are no significant differences in gender, sport dedication and type of sport (table 3).

Table 3: Descriptive and inferential statistics of dependence in regards to gender, age, weekly training sessions and sport type. 

Regarding the indicators of burnout, women show lower average scores than men in RSA (Z = 16,772.5; p = .002) and D (Z = 17,311.5; p = .006). On the other hand, older athletes exhibit higher mean scores in D (Z = 19,065.5; p = .001). As for dedication to sport, participants who spend more than 4 sessions a week on their workouts show a lower average score in RSA (Z = 19,972.5; p = .0001) but higher in E (Z = 20,474; p = .001). Statistically significant differences also appear depending on the modality of the sport (individual/ team) in RSA (Z= 16,776; p = .005), showing athletes who take part in team sports have a higher average range compared to athletes who participate on an individual basis (table 4).

Table 4: Descriptive and inferential statistics of dimensions of burnout in regards to gender, age, weekly training sessions and sport type. 

Concerning the relationship between exercise dependence and burnout, the results show that exercise dependence has a positive relationship (correlation significance .01) with the dimensions EFE and D, the highest (although low-moderate) corresponding to EFE and a low negative relationship with RSA (table 5).

Table 5: Correlations between exercise dependence indicators and dimensions of burnout. 

Discussion

The aim of the research was to analyse the prevalence and relationship between the symptoms of exercise dependence and burnout in athletes, as well as to investigate its relationship with other variables such as gender, age, sport dedication, and type of sport of the participants in the study.

The results show a higher percentage of dependence than those referred to by ^{Allegre, Souville, Therme and Griffiths (2006}) who found the risk of dependence was less than 5%. Well above the 2.7% recently found by ^{Zeulner, Ziemainz, Beyer, Hammon and Janka (2016}). However, other studies have found higher prevalence, placing the percentage at 13.6% in endurance athletes (^{Latorre, Jiménez, Párraga, & García, 2016}). ^{Marques, Peralta and Catuda (2015}) suggest that, although exercise dependence is a rare condition, it is triggered in professionals or semi-professionals related with sport between 26% and 34%.

Regarding the prevalence of the dimensions of burnout, ^{De Francisco et al. (2014}) found similar results when studying the prevalence of burnout through its three-dimensional conceptualisation, presenting in this study slightly more than one percentage point in RSA and less than one point (.7) for D. In general, few studies report percentages of the dimensions that make up the syndrome, with most studies providing data on burnout total or descriptive statistics for the three burnout dimensions.

Regarding the results found about the sociodemographic variables, no significant differences in exercise dependence were found in the same line as the recent study by ^{Reche et al. (2015}). ^{Allegre et al. (2007}) also found no differences regarding gender, highlighting this result as contrary to previous research where females showed higher values of dependence. These authors suggest that differences might not have been detected due to low representation of females in the sample (9%). Continuing with gender differences, for burnout, males show higher percentages for RSA and D, with no statistically significant differences in EFE. ^{Nafian et al. (2014}) point to gender role theories to explain the high levels of depersonalisation that the males showed in their sample because they tend not to show their emotions and to distance themselves, which could be applied to the results of this study.

As for the age of the athletes, results indicate that athletes under 18 years show a greater dependence on physical exercise. Similar data is presented in the study by ^{González-Cutre and Sicilia (2012}) in which the highest level of dependence is for young Spanish fitness centre users. Regarding the relationship between age and the dimensions of burnout, only significant results regarding D were found, where older athletes exhibit greater sport devaluation, which is consistent with the study by ^{Casagrande, Andrade, Viana and Vasconcellos (2014}), where older players showed higher averages in D. Meanwhile, ^{Pedrosa and García-Cueto (2014}) found higher levels of total burnout in adult athletes, and explained this on the grounds that greater demands in sports begin to appear when athletes leave adolescence.

As for sports dedication, it is not significantly associated with dependence exercise and therefore it is not proved to be joined with more training during the same day or more days of training, this is supported by ^{Kjelsås et al. (2003)} who do not consider high dedication to sport sufficient to create dependence exercise. Regarding burnout, studies indicate that more training demands result in higher levels of burnout (^{Gustafsson et al., 2007}; ^{Kellmann and Kallus, 2001}; ^{Kenttä et al., 2001}; ^{Silva, 1990}). ^{De Francisco et al. (2014}) indicate that most athletes, who train, show high levels of EFE, a fact that is corroborated in this study. It has not been the case for RSA where the highest values are for those who train for less than four hours and who therefore have a lower level of technique that in turn results in a lower sense of accomplishment. ^{Casagrande et al. (2014}) found results along this line, stating that athletes could be reducing the volume of training to feel they were not achieving the goals established.

The sport modality (individual or team) is not shown as a relevant variable in relation to dependence. ^{Reche et al. (2015}) came to the same conclusion. Sport modality do is relevant to burnout, specifically for RSA (athletes of team sports suffer more RSA). Previous studies indicate that it is more common in individual modalities (^{Coakley, 1992}; ^{Smith, 1986}) but also in collective modalities (^{Pedrosa & García-Cueto, 2014}), as it has been in this case. ^{Pedrosa and García-Cueto (2014)} suggest that collective athletes have a lower margin of influence on the results and successes to achieve than individual athletes.

Finally, it has been shown that exercise dependence relates positively with all dimensions of burnout, which requires a thorough study to analyse the specific role of dependence on the different dimensions of burnout, as the relationship is not solely due to training, since, if this were so, it would show significances only with EFE and not with the other dimensions of more psychological character. In addition, an inverse relationship with D is observed, and is explained by the fact that, with training, dependence and feelings of achievement are increased.

This research has several limitations. First, the sample size is relatively large, but also very heterogeneous. Second, an interaction between sociodemografic or sport variables and dependence or burnout could be present. Different studies (^{Cremades & Wiggins, 2008}; ^{Gustafsson et al., 2007}), showed an interaction that occurs between gender and type of sport over burnout. This research is a preliminary study so it would be interesting to investigate further and conduct analysis in this area since as indicated by ^{Judge, Bell, Theodore, Simon and Bellar (2012}): “women expressed more burnout than did men in some situations, but not others” (p. 236).