Study Design: Cross-sectional study. Objective: To evaluate isokinetic shoulder flexor-extensor (F/E) and abductor-adductor (Ab/Ad) torque ratios in individuals with paraplegia using a new interpretative approach. We proposed to study torque ratios according to joint angle sections (151 angle subgroups) over a selected range of motion. Setting: Pathokinesiology Laboratory, Montreal, Canada. Method: Sixteen individuals with complete motor paraplegia, without shoulder pain or impairment, were included in this study. After a preloading period of 1 s, maximum isokinetic concentric contractions of all muscle groups were completed at 30, 60 and 1201 s À1 over the entire tested ranges of motion (70 to À351 for the flexion-extension and 15 to 601 for the abduction-adduction). After the continuous torque curves were rebuilt, the mean F/E and Ab/Ad torque ratios were calculated and analyzed every 151. Results: A significant modification of the F/E (F ¼ 66.3; Po0.001) and Ab/Ad (F ¼ 100.6; Po0.001) torque ratios was observed according to the 151 angle subgroup evaluated. More precisely, a progressive decline of both the F/E and Ab/Ad ratios was noted as the shoulder flexion or abduction progressed. Angular velocity did not have any influence on torque ratio values. Conclusion: Angle subgroup torque ratio analysis leads to a better estimation of the balance between the agonist and antagonist muscle groups than does traditional peak torque ratio analysis. In individuals with paraplegia, this precise estimation of torque ratios may lead to the development of specific shoulder strengthening programs to prevent muscle imbalance and its consequences.