The lumbar multifidus (LM) plays a unique role for adjustment and support of lumbar lordosis. Although exercise therapies have shown that strengthening the multifidus produces reductions in low back pain, the quantification of the LMs role is undetermined. The goal of this in vitro study was to test the hypothesis that LM atrophy significantly reduces lumbar lordosis for spines in the upright posture.

The paraspinal muscles of six fresh human thoracolumbar spine (T12-pelvis) were surgically removed, leaving the spinal segments intact. The T12 level was fixed to a torso plate that facilitated anatomically appropriate trunk muscle attachment and dead weight simulation of torso weight (245 N). The pelvis was attached to a custom fixture that allowed pelvic tilt adjustment to the appropriate value for upright stance. In this experiment, three muscle groups were simulated using steel cables:​ the deep LM, erector spinae (ES), and rectus abdominus (RA). These cables were inserted at anatomically appropriate vertebral sites and were routed directly to digitally controlled torque motors. A total of five loading cases were investigated for three sets of experiments:​ neutral posture, axial rotation, and flexion/extension stability perturbation.

Results indicated a small change in lumbar lordosis for L4/L5 motion segment for atrophy while in the neutral posture. The data highlights how variations in muscle activation patterns alter lumbar lordosis and tissue stress distribution.