Predictors of excessive exercise in anorexia nervosa

Objective: The aim of the present study was to replicate and amalgamate findings from previous research into a comprehensive regression model predicting excessive exercise in individuals with anorexia nervosa (AN). Method: Participants were 153 patients admitted to an inpatient treatment program for AN. Excessive exercise status was defined as a minimum of 1 hour of obligatory exercise aimed at controlling shape and weight, 6 days per week in the month before admission. Results: Thirty-four percent (n = 52) of participants met criteria for excessive exercise. A logistic regression was conducted with excessive exercise status as the dependent variable and a number of variables previously found to be predictors of excessive exercise entered as independent variables on the same step. The overall regression model was statistically significant (P < .0005) and explained 31% of the variance in exercise status. Higher levels of dietary restraint (P = .03), depression (P = .04), and self-esteem (P = .02); lower levels of obsessive-compulsive symptomatology (P = .04); and the restricting subtype of AN (P = .03) were significantly associated with excessive exercise. Conclusions: Excessive exercise is associated with a number of independent psychologic and behavioral variables, some that suggest a negative impact and others that suggest positive effects. (C) 2010 Published by Elsevier Inc.