A novel testing protocol was used to investigate changes in neuromuscular performance, muscle recruitment, and spinal loading as subjects became fatigued while performing an isometric endurance test of varying torque requirements. There was decreased accuracy in maintenance of a reference torque but no change in response time as subjects became fatigued. The study of trunk-muscle recruitment indicated significant increases in internal oblique and latissimus dorsi muscle activity. This change in recruitment led to changes in spinal loading despite a relatively constant torque output. The use of an electromyogram (EMG)-assisted model demonstrated that when subjects are expected to become fatigued during test performance, the assumption of a constant maximal stress capacity of the muscle may not be robust.