Objective
To increase the accuracy and consistency in determining the level of radiculopathy by a needle electromyography (EMG) of multifidus muscle.

Method
We performed the EMGs on 29 patients with a low back pain to investigate an evidence of radiculopathy. All patients had the herniated nucleus pulposus (HNP) by a myelography, CT or MRI. The exclusion criteria were the patients with a scoliosis, spondylolisthesis or history of back surgery. We examined 5 points (P 1∼5) of the lumbosacral paraspinal muscles according to the paraspinal mapping by Haig et al and scored from 0∼4 according to the degree of abnormalities. The scores according to the points were correlated with the segments of radiculopathy and the levels of HNP.

Result: The maximal mean scores were 1.80⁑0.83 at P2 and 2.00⁑1.41 at P3 in a lumbar (L) 3, 4 radiculopathy, 2.00⁑0.56 at P5 in a L5 radiculopathy, and 2.13⁑0.64 at P4 and 2.63⁑0.51 at P5 in a S1 radiculopathy. The sensitivity/specificity was high at P2, P3 in a L3, 4 radiculopathy, at P4 in a L5 radiculopathy, at P5 in a S1 radiculopathy.

Conclusion
The results suggest that the localization of lumbosacral radiculopathy by a needle EMG of multifidus muscles provides an easy accessibility, better accuracy and consistency to determine the level of radiculopathy.