P1-latency interroot comparison enhances the validity of scalp-recorded somatosensory-evoked potentials to diagnose nerve root dysfunction in sciatica

• Published in: Spine (Philadelphia, Pa. 1976) Vol. 22; no. 7; p. 798
• Main Authors: Pape, E, Eldevik, O P
• Format: Journal Article
• Language: English
• Published: United States 01.04.1997
• Subjects: Adolescent; Adult; Aged; Evoked Potentials, Somatosensory; Female; Humans; Lumbosacral Region - diagnostic imaging; Lumbosacral Region - physiopathology; Male; Middle Aged; Myelography - standards; Nerve Compression Syndromes - diagnostic imaging; Nerve Compression Syndromes - physiopathology; Polyradiculoneuropathy - diagnostic imaging; Polyradiculoneuropathy - physiopathology; Prospective Studies; Reference Standards; Reproducibility of Results; Scalp; Sciatica - diagnostic imaging; Sciatica - physiopathology
• ISSN: 0362-2436
• DOI: 10.1097/00007632-199704010-00019

A prospective validity study was done of scalp-recorded somatosensory-evoked potentials as a diagnostic procedure to show lumbosacral radiculopathy in 100 consecutive patients with unilateral or bilateral sciatica. To determine the validity gained by the use of P1-latency interroot comparison to show P1-latency prolongation. The validity of scalp-recorded somatosensory-evoked potentials in diagnosing lumbosacral radiculopathy has been debated and is uncertain. Sensory nerves representing nerve roots L4, L5, and S1 were stimulated bilaterally. Height-corrected P1-latency, two new P1-interroot comparison-based criteria, and absence of P1 were studied. The gold standard was defined as clinically-involved nerve roots with radiologic nerve root compression. The false-positive nerve root compression rate was determined, and the gold standard was corrected accordingly. Clinically relevant cut-off values were defined by multilevel likelihood ratio analysis. The positive and negative likelihood ratios of P1-latency prolongation were 6.79 and 0.53, respectively, for the gold standard, and 10.57 and 0.21, respectively, for the corrected gold standard. The validity was not reduced when scalp-recorded, somatosensory-evoked potentials were blinded to the radiologic results. Absence of P1 was associated to the gold standard and the corrected gold standard. Compared with the combined use of P1-latency interside difference and height-corrected latency, the combination of P1-interroot comparison and height-corrected P1-latency increased the sensitivity by 20% for the gold standard and 32% for the corrected gold standard, and when absent P1 was added, the overall sensitivity was 53% for the gold standard and 81% for the corrected gold standard. The corresponding specificity was 92%, and in asymptomatic nerve roots it was 98%. The P1-interroot comparison permits the use of scalp-recorded somatosensory-evoked potentials as a contributory "rule in" procedure in patients with sciatica.