Electromyogram diagnosed chronic lumbar radiculopathy is highly associated with L4-S1 regional multifidus atrophy on magnetic resonance imaging.

backgroundLumbar multifidus atrophy (LMA) is recognised as having clinically significant implications. Pathology producing lumbar radiculopathy (LR) is known to be associated with LMA, but there is a paucity of large studies formally evaluating the relationship between LR electrodiagnostics and LMA. Our aim was to assess the association of electrodiagnostically proven LR with the presence of LMA on Magnetic resonance imaging (MRI).

methodsA cohort of 100 consecutive eligible patients with chronic lumbar spinal symptoms was analysed from 119 patients who underwent electrodiagnostic evaluation. Electrodiagnostic assessment included nerve conduction studies (NCS) and needle electromyography (EMG). NCS was used to exclude peripheral neuropathy and other abnormalities, while the diagnosis of chronic lumbar radiculopathy (CLR) was based on needle EMG findings and overall electrophysiological assessment. Patients with EMG-confirmed CLR were assigned to the positive LR cohort, while those with EMG-negative findings for LR served as the control cohort. LMA was assessed on MRI by two independent observers using 3D proton density weighted gradient echo sequences. LMA was graded according to the modified Goutallier classification and quantified using the percentage of fatty infiltration derived from the Dixon dual echo method 4. A threshold of > 15% fatty infiltration or modified Goutallier grade ≥ 1 was designated as atrophy. These imaging findings were analysed for an association with EMG-confirmed radiculopathy.

resultsOne hundred eligible positive CLR patients were analysed. Modified Goutallier grading demonstrated 2% grade 0, 68% grade 1, 19% grade 2 and 11% grade 3 LMA. Ninety-eight per cent of patients with EMG-proven LR had LMA, compared with 64% of symptomatic patients with normal EMG findings (Fisher's exact test, p < 0.001; odds ratio 27.2). In the control cohort, 36% demonstrated grade 0, 43% grade 1, and 21% grade 2 LMA, with no patients demonstrating grade 3 atrophy. The prevalence of LMA was significantly higher in the CLR cohort compared with symptomatic patients with normal EMG findings.

conclusionsThis diagnostic study reveals EMG-proven CLR is highly associated with MRI-based LMA. Based on these findings, LMA, although not pathognomonic and commonly observed in symptomatic patients, may represent a potential biomarker for CLR. CLINICAL TRIAL NUMBER: Not applicable.