Persons with bulimia nervosa (BN) or binge eating disorder (BED) have an elevated risk of non-communicable diseases (NCDs). However, lowering this risk is rarely addressed in standard cognitive-behavioral treatment (CBT). We aimed to compare CBT with an intervention combining physical exercise and dietary therapy (PED-t), and hypothesized that the PED-t would do better than CBT in lowering the risk of NCD both initially and longitudinally. In this study, 164 women with bulimia nervosa or binge eating disorder were randomly assigned to 16-weeks of outpatient group therapy with either PED-t or CBT. Body composition (BC) was measured by dual-energy X-ray absorptiometry. Measures of physical fitness (VO₂peak and one repetition maximum (1RM) in squats, bench press, and seated row) were also recorded. All measurements were completed baseline, post-treatment, and at 6- and 12-month follow-ups, respectively. Our results showed that PED-t improved more than CBT on mean (99% CI) absolute Vo2peak; 57,2 (84.4, 198.8) mL ( = 0.22, < 0.001) post-treatment. There were small to medium long-term differences in 1RM after PED-t compared to CBT. BC deteriorated in both groups during follow-up. Neither the PED-t nor the CBT lowered the risk for NCDs. Clearly, other approaches need to be considered to promote physical fitness and lower the risk of NCDs among individuals with BN and BED.