In elite climbers, the development of "climber's back" has often been subjectively referred to. However no scientific proof is present. In a cross-sectional cohort study, the spines of 80 healthy asymptomatic male sport climbers were examined. The performance-oriented sport climbers (SC) trained regularly (9.8 +/- 4.3 hr/wk) and had a mean climbing ability of 9.7 +/- 0.6; the 34 recreational climbers (control group) (RC) climbed less frequently (3.4 +/- 2.0 hr/wk) and had a mean climbing ability of 6.0 +/- 0.9. Measurement of the sagittal thoracolumbar spine was performed using the "SpinalMouse". The kyphosis angle in the erect posture was significantly greater in SC verses the RC. The lordosis angle was also greater in SC versus RC but did not reach significance. No significant differences were found in flexion and extension. Further evaluation of the SC group was carried out by subdividing them to a moderate group (SC-moderate) (n = 17) and top-level group SC-top level) (n = 29). Here the kyphosis angle was significantly greater in SC-top-level than in SC-moderate. The results demonstrated that "climber's back" was characterized by an increased thoracic kyphosis, increased lumbar lordosis, and was probably influenced by shortened pectoralis muscles. The climbing ability level was strongly correlated to the postural adaptations.