Table 3.

Outcome measures and results of included studies

Authors (Year) (Country)	Outcomes	Quality index assessment	Results relating to OCD	Results relating to OCPD	Limitations
Anderluh et al. (2009) (Slovenia/UK)[35]	EATATE1 to assess diagnosis, screen for lifetime obsessive compulsive disorder and eating disorder diagnosis using ICD-10 criteria; as well as obsessive compulsive traits in childhood.	9	Groups did not differ in lifetime duration of excessive exercising. No differences between groups in frequency of current or lifetime OCD.	No differences between groups in frequency of current OCPD. Children who were rule bound/cautious exercised excessively later in life (p < .005, p < .02).	Retrospective assessment subject to memory biases, although anchor points were used. Data from informants could have assisted with this. Participants recruited from secondary and tertiary treatment centre.
Bewell-Weiss & Carter (2010) (Canada)[36]	EDE2; EDE-Q; BSI for anxiety; BDI-II for depression; RSES for self-esteem; Padua Inventory for obsessive compulsive symptoms; EDI for eating disorder attitudes and behaviours.	11	Overall model significant (p < .05) explained 31% of variance in exercise status. Restraint, self esteem and depression positively associated with exercise (p < .05), OC symptomatology negatively associated with exercise status (p = .038).	NA	Need to replicate findings with other measures. Study of factors associated with specific definition of exercise. Cross-sectional nature could not demonstrate direction of associations. No measure of OCPD.
Davis & Claridge (1998) (Canada/UK)[37]	EPQR3 for addictiveness; obsessive-compulsive personality subscales; Drive for thinness for weight preoccupation; Interview to determine lifetime and current exercise status- classified as excessive or non-excessive exercisers.	11	NA	Both addictiveness and obsessive-compulsiveness were positively associated with over-exercising (both currently and historically, p < .05 and p < .01 respectively).	Patients were specifically chosen to represent the two diagnoses, although commonly both AN and BN features co-occur in clinical syndromes and in personality structure of patients.
Davis & Kaptein (2006) (Canada)[38]	Interview to determine lifetime and current exercise status (excessive or non-excessive); MOCI4 to assess for OCD symptomatology; obsessive-compulsive personality subscales; BMI.	11	Excessive exercisers showed higher intensity/number of OCD symptoms than non-excessive patients (p = .007). There was a decline in symptom severity between admission and discharge (p < .001).	Excessive exercisers demonstrated greater OC personality traits than non-excessive patients (p = .03). There was no significant decline in OC personality traits between admission and discharge.	Self-report recall data was used in this study. Indirect historical data is necessary, on account of low prevalence of AN-R. Difficulties with prospective designs.
Davis et al. (1998) (Canada)[2]	MOCI5 -symptoms of OCD; Obsessive Compulsive Personality subscales; MPS for perfectionism; CES for commitment to exercise; EDI: weight preoccupation; BES for body image; JFFIS for self-esteem; BMI.	11	Exercisers scored significantly higher than non-exercisers on OC symptomatology, (p = .02). Exercisers also reported more obligatory and pathological attitudes to exercise (p < 0.01).	Exercisers scored significantly higher than non-exercisers on OC personality characteristics (p < .05), and self-oriented perfectionism (p < .05).	No information as to whether exercise and obsessionality influence prognosis. Obsessionality data obtained solely from self-report data, not structured diagnostic interview.
Davis et al. (1995) (Canada)[39]	SCL-906 to measure obsessive-compulsiveness; Drive for thinness to measure weight preoccupation; CES to measure commitment to exercise; interview to assess for physical activity.	12	Obsessive-compulsiveness significantly positively related to level of activity among AN patients (p < .01), and obligatory and pathological aspects of exercise were related to weight preoccupation (p < .01).	NA	Proposed activity-based anorexia model explains AN development only for some individuals. Does not take into account motivational factors, differences in selecting forms of exercise and reasons for exercising. No OCPD measure.
Holtkamp et al. (2004) (Germany)[40]	SIAB7 to assess AN subtype; SCL-90-R to assess for anxiety, depression and obsessive-compulsiveness.	11	Obsessive-compulsiveness was not associated with physical activity levels (r = -.072, p = .705). Regression model based on BMI, food restriction, subtype, anxiety, depression and obsessive-compulsiveness (OC) explained 64% of variance in model, OC was not a significant contributor.	NA	Small sample size; data on food restriction were answered retrospectively; need more detailed measure of OCD symptoms; need to examine OCPD symptoms; SCL-90-R only validated for people 14 years and older.
Naylor et al. (2011) (UK)[41]	Exercise Frequency; CET8 to measure beliefs about exercise; OBQ-44 to assess OCD constructs; OCI-R to assess distress associated with OCD symptoms; EDE-Q.	12	Clinical: women with higher exercise beliefs had higher levels of obsessive beliefs and obsessive compulsive behaviours (p < 0.01). OBQ and OCI-R accounted for significant increase in variance of weight control exercise explained.	NA	Cross-sectional design of study does not show direction of variables; rather just associations. Use of Self-report measures risks of socially desirable responses. Student samples are arguably unrepresentative of general population, could use other control groups.
Penas-Lledo et al. (2002) (Spain)[31]	EAT-409: overall level of eating pathology. BITE: bulimic attitudes and behaviours. SCL-90-R: current psychological symptoms BMI.	11	AN patients who exercised had higher levels of eating pathology (EAT; p < .01) Exercisers had higher levels of anxiety and depression on SCL-90-R (p < .01), but not OCD symptoms (p > .05).	NA	Problems in definition of excessive exercise used to classify groups; require different measures to examine OCD symptoms in more detail; no measurement of OCPD symptoms
Shroff et al. (2006) (USA)[42]	Clinical variables: ED duration, current/minimum/maximum BMI obtained; SIAB; SCID for diagnosis; TCI for temperament; MPS for perfectionism; STAI for anxiety; Y-BOCS for OC symptoms; YBC-EDS; excessive exercise classification from SIAB.	13	Excessive exercise was associated with greater severity of ED symptoms, worst ritual, preoccupation and worst motivation to change in YBC-EDS. Also associated with higher obsessions and compulsions (YBOCS) (p < .001)	Excessive exercise was associated with all perfectionist traits (p < .001), as measured by MPS.	Exercise group determined by retrospective reports of exercise behaviour. Exercise assessment not comprehensive. Unable to determine association between duration of excessive exercise and other ED behaviours.

¹EATATE EATATE Lifetime Diagnostic Interview [43].

²EDE Eating Disorder Examination [44]; EDE-Q Eating Disorder Examination-Questionnaire [45]; BSI Brief Symptom Inventory [46]; BDI-II Beck Depression Inventory [47].

RSES Rosenberg Self-Esteem Scale [48] Padua [49]EDI Eating Disorder Inventory [50].

³Eysenck Personality Questionnaire-Revised [51];Obsessive-compulsive personality subscales [52]; Subscale from the Eating Disorder Inventory [50]; exercise interview as used previously by [21,30].

⁴MOCI Maudsley Obsessive-Compulsive Inventory [53]; personality subscales [52].

⁵MOCI Maudsley Obsessive-Compulsive Inventory [53]; personality subscales [52]; MPS Multidimensional Perfectionism Scale [54]; CES Commitment to Exercise Scale [55]; EDI Eating Disorder Inventory [50]; BES Body Esteem Scale [56]; JFIS Janis-Field Feelings of Inadequacy Scale [57].

⁶SCL-90 Symptom Checklist-90 [58]; Drive for thinness from EDI [50]; CES [55].

⁷SIAB Structured Interview of Anorexia and Bulimia Nervosa [59]; SCL-90-R [60].

⁸CET Compulsive Exercise Test [32]; OBQ-44 Obsessive Beliefs Questionnaire-44 [61]OCI-R Obsessive Compulsive Inventory-Revised [62]; EDE-Q Eating Disorder Examination-Questionnaire [63].

⁹EAT-40 Eating Attitudes Test [64]; BITE Bulimic Investigatory Test [65]; SCL-90-R [66]; SIAB Structured Interview for Anorexia and Bulimia Nervosa [59]; SCID [67]; TCI Temperament and Character Inventory [68]; Frost MPS [54]; STAI [69] Y-BOCS [70] YBC-EDS [71].