Background and aims: Exercise addiction receives substantial attention in the field of behavioral addictions. It is a unique form of addiction because in contrast to other addictive disorders it is carried out with major physical-effort and high energy expenditure. Methods: A critical literature review was performed. Results: The literature evaluation shows that most published accounts report the levels of risk for exercise addiction rather than actual cases or morbidities. The inconsistent prevalence of exercise addiction, ranging from 0.3% to 77.0%, reported in the literature may be ascribed to incomplete conceptual models for the morbidity. Current explanations of exercise addiction may suggest that the disorder is progressive from healthy to unhealthy exercise pattern. This approach drives research into the wrong direction. Discussion: An interactional model is offered accounting for the adoption, maintenance, and transformation of exercise behavior. The here proposed model has an idiosyncratic black-box containing the antecedents and characteristics that are unique to the individual, which cannot be researched via the nomothetic approach. Subjective aspects in the black-box interact with stressful life events that force the person to cope. The range of coping may be wide. Escape into exercise depends on personal (subjective) and situational (objective) factors, but the subjective components are inaccessible for a priori scholastic scrutiny. It is our view that currently only this dual interactional model may account for the fact that exercise addiction emerges suddenly and only in a few individuals from among those at high risk, estimated to be around 3.0% of the exercising population.

Background and aims Cited in over 100 articles, the interactional model of exercise addiction (Egorov & Szabo, 2013) forms the theoretical foundation of many studies on the risk of exercise addiction. Still, the inclusion of previously omitted determinants could make it more useful. Therefore, this review presents the expanded version of the original model. Method We added ‘self-concept’ as another determinant in the ‘personal factors’ domain and ‘attractive alternatives’ to the ‘situational factors’ domain. Further, we doubled the reasons for exercise in the ‘incentives for exercise domain.’ Last, we added a new domain, the ‘exercise-related stressors,’ to illustrate that exercise itself might be a source of stress. Results The expanded model is more inclusive and accounts for a greater combination of interactions playing roles in exercise addiction. Overlooking the eventuality that stress resulting from exercise might also fuel the dysfunction was a significant omission from the original model, rectified in the current update. Finally, the new expansions make the model more applicable to competitive situations too Conclusion The expanded interactional model of exercise addiction is more comprehensive than its original version. It also accounts for the exercise or sport-related stress as possible fuel in addictive exercise behavior.

Recent research suggests that the risk of exercise addiction (REA) is primarily predicted by passion. However, this relationship stems from cross-sectional studies. The objective of this 12-week longitudinal investigation was to examine and compare the temporal changes in REA and passion among women and men (N = 149) who just started a new sport activity. To further test their similarities and distinctiveness, we also examined how exercise motivations predict REA and passion. Latent growth modeling showed that REA and passion were high at baseline and showed a slight increase over the 12 weeks. Passion was predicted by several motivational factors, whereas REA was predicted by being female, team sport participation, higher exercise intensity, and identified motivation. These findings suggest that the development of passion and REA over time, both being associated with sport motivations, manifest independent patterns. Consequently, despite the reported strong relationship between the two, they appear to be independent constructs.

No abstract