METHODS Test-retest study. OBJECTIVE To describe the test-retest reliability of an individualized approach for peak oxygen uptake (VO2peak) testing in a clinical context, accommodating inpatients with heterogeneous levels of physical functioning. METHODS Inpatient rehabilitation. METHODS Test-retest on two types of devices (recumbent stepper and arm-crank ergometer) with two predefined protocols of varying load increments on each device. No participants performed all four test modalities. VO2peak, respiratory exchange ratio (RER), peak workload, and the reason for termination were reported. Of 23 participants included, 21 completed the test-retest procedure. Participants were inpatients, ≥18 years, who sustained a spinal cord injury (SCI) within the last 12 months at the time of admission. The median age was 52 years and 74% were male. Median time since injury at the time of the first test (T1) was 5.5 months. Seventeen were classified as American Spinal Injury Association Impairment Scale D. Intraclass correlation coefficients (ICC) and Bland-Altman plots for VO2peak was calculated to determine intra- and intertester reliability of VO2peak. RESULTS Mean VO2peak obtained at T1 was 1.91 L/min (95%CI: 1.31-2.51); 1.26 L/min and 2.69 L/min for the recumbent stepper protocols; 0.61 L/min and 1.39 L/min for the arm-crank ergometer protocols. ICC for intra- and intertester reliability was 0.997 (95%CI: 0.986-0.999) and 0.994 (95%CI: 0.978-0.998) respectively. Bland-Altman plots showed a mean difference of -0.005 (SD 0,12) for overall test-retest. CONCLUSIONS Test-retest reliability was high. Mean VO2peak at discharge from rehabilitation was higher than previously reported and individualized test modalities seemed to yield more optimal results in a heterogenous and representative sample during rehabilitation.