Consensus has yet to be achieved on whether obesity is inexorably tied to poor fitness. We tested the hypothesis that appropriate reference of cardiopulmonary exercise testing (CPET) variables to lean body mass (LBM) would eliminate differences in fitness between high-BMI (≥ 95th percentile, n = 72, 50% female) and normal-BMI (< 85th percentile, n = 142, 49% female), otherwise-healthy children and adolescents typically seen when referencing body weight. We measured body composition with dual x-ray absorptiometry (DXA) and CPET variables from cycle ergometry using both peak values and submaximal exercise slopes (peak VO2, ΔVO2/ΔHR, ΔWR/ΔHR, ΔVO2/ΔWR, and ΔVE/ΔVCO2). In contrast to our hypothesis, referencing to LBM tended to lessen, but did not eliminate, the differences (peak VO2 [p < .004] and ΔVO2/ΔHR [p < .02]) in males and females; ΔWR/ΔHR differed between the two groups in females (p = .041) but not males (p = .1). The mean percent predicted values for all CPET variables were below 100% in the high-BMI group. The pattern of CPET abnormalities suggested a pervasive impairment of O2 delivery in the high-BMI group (ΔVO2/ΔWR was in fact highest in normal-BMI males). Tailoring lifestyle interventions to the specific fitness capabilities of each child (personalized exercise medicine) may be one of the ways to stem what has been an intractable epidemic.