Our study findings showed that the DOS scale was poorly able to detect OCD at both cutoff points (25 and 30), with low PPV and high NPV values. The sensitivity values at both cutoff points were not adequate, while we obtained high specificity. The AUC value was 0.7, indicating a poor ability of the DOS scale to predict OCD symptoms [38]. These results consolidate the findings of previous authors that ON and OCD are separate phenomena [39, 40] and that ON shares more common features with disordered eating rather than OCD [26, 27]. While there is no solid evidence about the correlation of ON with OCD, further research remain warranted to resolve this controversy. A review on psychosocial variables associated with ON [19] reported that perfectionism, orthorexia traits, psychopathology, disordered eating and antecedants, dieting, negative body image and drive for thinness were associated with higher ON. Barthels et al. [41] suggested that ON and ED share common psychopathological characteristics such as low body acceptance, which links both disorders together.
The results in the regression analyses highlighted that higher OCD and EAT total scores (more inappropriate eating) were significantly associated with more ON tendencies and behaviors. As mentioned in the introduction, ON is characterized by obsessions and compulsions about the quality of food [42]. Previous studies concluded that more OCD symptoms are linked with more ON behaviors [2, 43,44,45]. In particular, the OCI washing subscale was found to be significantly associated with ON; we hypothesize that this finding might be related to the fact that during the COVID-19 pandemic, handwashing is applied as an essential measure to prevent the disease’s spread [46, 47], since hands are a known vector in the transmission of microorganisms. During the pandemic, there has been extensive efforts implemented to raise awareness towards handwashing as recommended by the World Health Organization and other healthcare institutions to stop the COVID-19 virus spread. Our findings showed a close association between EAT and ON, which can be demonstrated by the fact that EAT assesses disorderd eating habits together with the social pressure, food pre- occupation, purging behaviors and food awareness, which are also able to predict orthorexic behaviors [48]. Furthermore, ON share similarity with eating disorders where both have impulsive persistent healthy food thoughts.
Higher physical activity index was significantly associated with more ON, consolidating the findings of previous papers [18, 43] that found a significant close relation between ON symptomatology and physical activity. A recent systematic review reinforced this relationship and showed a slight correlation with exercise and moderately with exercise addiction, conveying a shared variance between those two conducts [49]. Our results can be explained by the fact that orthorexic people strictly exercise to promote their health as engagement in regular physical activity plays a vital role in a healthy lifestyle, weight and stress management [50].
Clinical implications
Studying the relation between psychological disorders and ON has a potential advantage to highlight treatment recommendations and preventive guidelines applications. In addition, understanding the impact of personality traits along with eating disorders on ON individuals improves prediction about the risk of ON manifestations. Identification of eating disorders features by enhancing patient education about the various etiologic factors (physical activity, perception of food quality and quantity) can be translated into decreased progression of ON.
Limitations
This is a cross-sectional study where cause-effect relationship and causality of the relationship between OCD and ON cannot be established. The study took into account many factors associated with ON but other factors might not be taken into consideration, which introduced the risk of residual confounding. Symptoms were self-reported and not evaluated by a clinician. Data was collected online, excluding those who do not have Internet access or a smartphone, predisposing us to a selection bias. The majority of the participants was females and with a low mean age, thus, the results cannot be generalized to the whole population. Information bias is also present since participants might give wrong information in all cross-sectional studies. The design is limited due to the use of self-report assessment (and not a psychiatrist diagnosis), internet-based tools to validate an instrument and the lack of a control group. Finally, there is no valid and coherent gold standard measure to define which participants are classified as ON and which are not.