Background: A randomized trial of a pedometer-based intervention with weekly activity goals led to a modest increase in step count among dialysis patients. In a secondary analysis, we investigated the effect of this intervention on body composition.Methods: Sixty dialysis patients were randomized to standard care or a 6-month program consisting of 3 months of pedometers and weekly step count targets and 3 months of post-intervention follow-up. We obtained bioelectrical impedance spectroscopy (BIS) data on 54 of these patients (28 control, 26 intervention) and used linear mixedmodeling (adjusted for sex and dialysis modality) to estimate differences in change in total-body muscle mass (TBMM) adjusted for height 2 , fat mass (kg), and body mass index (BMI) (kg/m 2 ) between control and intervention groups.Results: The median age of participants was 57.5 years (53-66), and 76% were men. At baseline, there was no significant difference between groups in age, BMI, race, or body composition, but there were more men in the intervention group. After 3 months, patients in the intervention group increased their average daily steps by 2414 (95% CI 1047, 3782) more than controls (p < 0.001), but there were no significant differences in body composition. However, at 6 months, participants in the intervention had a significantly greater increase from baseline in TBMM of 0.7 kg/m 2 (95% CI 0.3, 1.13), decrease in fat mass (− 4.3 kg [95% CI -7.1, − 1.5]) and decrease in BMI (− 1.0 kg/m 2 [95% CI -1.8, − 0.2]) relative to controls. In post-hoc analysis, each increase of 1000 steps from 0 to 3 months was associated with a 0.3 kg decrease in fat mass (95% CI 0.05, 0.5) from 0 to 6 months, but there was no dose-response relationship with TBMM/ht 2 or BMI. Conclusion: A pedometer-based intervention resulted in greater decreases in fat mass with relative preservation of muscle mass, leading to a greater decrease in BMI over time compared with patients not in the intervention. These differences were driven as much by worsening in the control group as by improvement in the intervention group.Step counts had a doseresponse relationship with decrease in fat mass.Trial registration: ClinicalTrials.gov (NCT02623348). 02 December 2015.