Predictors of Resumption of Menses in Anorexia Nervosa: A 4-Year Longitudinal Study

Objective Amenorrhea is a disabling medical consequence of anorexia nervosa (AN); therefore, resumption of menses (ROM) represents an important goal in the treatment for these patients. The aim of the present study was to evaluate possible clinical, psychopathological, and biological predictors of ROM, including age, body mass index (BMI), AN subtype, childhood abuse, duration of illness, general and eating disorder (ED)-specific psychopathology, and sex hormones. Methods Fifty amenorrheic patients with AN were enrolled. Baseline clinical data and information on childhood abuse were collected. Questionnaires to evaluate general and ED-specific psychopathology were administered, and blood samples were drawn. All patients received treatment as usual and underwent regular follow-up visits for 4 years or until ROM. Time to ROM, BMI at last evaluation, and data regarding diagnostic crossover into bulimia nervosa were collected. Results Twenty-nine (58.0%) patients recovered menses. Diagnostic crossover was associated with a higher probability of ROM (odds ratio = 10.3,p= .030). Time-to-event analysis showed that a shorter duration of illness (chi(2)(1) = 11.00,p= .001), binge-eating/purging subtype (chi(2)(1) = 7.01,p= .008), and history of childhood abuse (chi(2)(1) = 4.03,p= .045) were associated with an earlier ROM. Furthermore, higher baseline ED-specific psychopathology was associated with a reduced likelihood for ROM, whereas higher general psychopathology and follicle-stimulating hormone levels predicted an earlier ROM (all,p< .050). Age, BMI, luteinizing hormone, and estrogen hematic levels had no predictive value with respect to ROM. Conclusions The present study provides data in support of an integrated model, emphasizing the importance of duration of illness, childhood abuse, and psychopathological characteristics of amenorrheic patients with AN in predicting ROM.