PurposeThe purpose of the present study was to determine the reliability of the exercise response (predicted peak VO2) using the total body recumbent stepper (TBRS) submaximal exercise test in: 1) healthy adults 20–70 years of age and 2) adults participating in inpatient stroke rehabilitation. We hypothesized that the predicted peak VO2 (Visit 1) would have an excellent relationship (r > 0.80) to predicted peak VO2 (Visit 2). We also wanted to test whether the exercise response at Visit 1 and Visit 2 would be significantly different.MethodsHealthy adults were recruited from the Kansas City metro area. Stroke participants were recruited during their inpatient rehabilitation stay. Eligible participants completed 2 TBRS submaximal exercise tests between 24 hours and 5 days at similar times of day.ResultsA total of 70 participants completed the study. Healthy adults (n = 50) were 36 M, 38.1 ± 10.1 years and stroke participants (n = 20) were 15 M, 62.5 ± 11.8 years of age. The exercise response was reliable for healthy adults (r = 0.980, p<0.01) and stroke participants (r = 0.987, p<0.01) between Visit 1 and Visit 2. Repeated Measures ANOVA showed a significant difference in predicted values between the two visits for healthy adults (47.2 ± 8.4 vs 47.7 ± 8.5 mL∙kg-1∙min-1; p = 0.04) but not for stroke participants (25.0 ± 9.9 vs 25.3 ± 11.4 mL∙kg-1∙min-1; p = 0.65).ConclusionThese results suggest that the exercise response is reliable using the TBRS submaximal exercise test in this cohort of healthy adults and stroke participants.