A randomized controlled trial of physical exercise- and dietary therapy versus cognitive behavior therapy: Treatment effects for women with bulimia nervosa or binge eating disorder

Background: Eating disorders (EDs) are among the top ten of the gender and age adjusted global burden of diseases in terms of poor quality of life, affecting young women in particular. Less than half of the persons with bulimia nervosa (BN) or binge eating disorder (BED) are detected and offered treatment for their ED in primary care. Besides low detection rate; low mental health literacy, and long waitlists for special care are important causes to this scenario. Cognitive behavior therapy (CBT) is recognized as the preferred evidence based treatment option for BN and BED, still more than 60% do not fully abstain from symptoms. There is a need to explore new treatment options that circumvent the challenges with low treatment access and poor remission rate. Evidence suggest that regular physical activity effectively prevents and treats physical- and mental morbidity and mortality, contributing to improvements in quality of life. Physical activity is however, rarely incorporated in treatment of EDs out of fear of exacerbating the compulsive and excessive nature of exercise in patients for compensatory or affect regulation purposes. Objectives: To evaluate the effect of a new treatment method for women with BN or BED, combining guided physical exercise and dietary therapy (PED-t), being offered as group therapy. The novel treatment method was compared to the effect of cognitive behavior therapy (CBT), and a waitlist control group. The first paper describes the rationale for, and the specific study protocol from the PED-t trial. The second paper describes the physical fitness in women with BN or BED more thoroughly than previously in the literature. In the third paper we investigated the effect from PED-t or CBT on compulsive exercise and level of physical activity. In the fourth paper we investigated the effect from PED-t or CBT on remission from ED, ED-symptomology, and measures of mood and quality of life. Methods: During 2014-2016 totally 187 women with BN or BED, aged 18-40 and with BMI 17.5-35 were enrolled in this RCT, and allocated to PED-t (n=82) or CBT (n=82), or temporarily placed in a waitlist control group (n=23). Effect from 16 weeks of treatment by either CBT or PED-t, or being in control group, was evaluated and compared at baseline (T1), post-test (T2) and follow-up periods (6 months, T3, and 12 months, T4). Outcomes were blood pressure, cardiorespiratory fitness (CRF), muscle strength (1RM), physical activity, body composition, compulsive exercise (CE), remission from diagnosis, and alleviation of ED-symptoms (by EDE-q) and comorbidity. Measures were by cardiopulmonary exercise testing, 1RM strength tests, DXA, objective registration of physical activity, and questionnaires. Results: In total 156 met for baseline screening, of whom 103 were diagnosed with BN and 53 with BED. Overall, participants with BN or BED displayed adequate physical fitness; however, a high number had high blood pressure, low CRF and unfavorable body composition. The number of randomized participants (n=164) that met for therapy was 149, of whom 112 completed treatment (32% drop out). Dropouts and completers were different by a lower mean (CI95) score for depression amongst completers (-3.08 -5.95, -0.21, g=0.39, p=0.035), and significantly more from CBT were lost to follow-up at T3 and T4 compared to PED-t. About 40-70% of all participants scored above clinical cut-off in the compulsive exercise test (CET) at baseline. CBT and PED-t were equally effective in reducing compulsive exercise after 16 weeks of treatment (P < 0.01, Hedges g ~ 0.4), with sustained long-term effects (T3-T4). The proportion of participants that complied with the official recommendation for physical activity (~47%) neither changed following treatment, nor emerged different between the therapy arms. After treatment mean EDE-q global score improved more in the PED-t group compared to the CBT group (-0.66, [CI99 -1.23, -0.1], g=0.52, p <0.003) and to the control group (-1.15, [CI99 -1.97, -0.34], g=1.00, p<0.001), whereas CBT did not differ from the control group (-0.49 [CI99-1.32, 0.34], g=0.48, p=0.12). Numbers in full- or partial remission were higher in PED-t (29.0% and 19.7%) and CBT (12.4% and 16.7%) compared to control (0.1% and 5.6%), p