EMG beats MRI for diagnosing low back pain, spinal stenosis

Needle electromyography (EMG) is a better (and much cheaper) alternative to magnetic resonance imaging (MRI) for patients with low back pain (LBP) and might both improve diagnosis and prevent unnecessary surgery, Dr Andrew J Haig reports in the December 1, 2005 issue of Spine Haig and colleagues at the University of Michigan (Ann Arbor) conducted a rigorous prospective, masked, double-controlled diagnostic trial in patients whose MRIs suggested back pain or spinal stenosis and found that EMG had “statistically significant, clinically meaningful specificity for spinal stenosis and detects neuromuscular diseases that may masquerade as stenosis.”

“If you want to find out if the patient’s complaint relates to ‘spinal stenosis,’ get an EMG. It won’t be perfectly sensitive, but a positive EMG tells you that there is spinal nerve involvement. To help eliminate neuropathy, peripheral nerve entrapment, tumors of the pelvis, and many spinal tumors, it’s EMG again,” Haig told rheumawire. “MRI might detect the rare spinal tumor that is missed on EMG and is necessary before spinal surgery or injection, but clinicians should not even look at the MRI films until they’ve made a clinical impression.” First blinded, controlled study of EMGThis study, which was the first blinded or controlled study of EMG for any disorder, requires replication, but Haig predicts, “I think the day will come when we order an EMG, erythrocyte sedimentation rate, and a plain X-ray, nothing else unless surgery or injection is planned.” The investigators enrolled 150 subjects, including 30 asymptomatic volunteers, 30 with LBP but no MRI evidence of spinal stenosis, and 90 with varying severity of spinal stenosis. “This paper was designed to meet the strictest criteria of evidence-based medicine, so it included only persons with very clear stenosis and persons who were very clearly without stenosis,” Haig said.All subjects filled out a number of questionnaires, including the Pain Disability Index, the Quebec Back Pain Disability Index, the McGill pain scale, a visual analog pain scale, and a pain drawing, plus a five-page clinical spine questionnaire. Each subject performed a 1.5-minute-walk test and wore a pedometer at home for one week. All subjects had noncontrast lumbosacral spinal MRI scans, and all MRI scans underwent blinded interpretation, including measurements at each spinal level, clinical impression at each spinal level, and overall clinical impression. Next, all subjects underwent detailed EMG using a 50- to 75-mm monopolar needle. This included exploration of five muscles with overlapping root innervation: tensor fascia lata, vastus medialis, tibialis anterior, extensor hallucis longus, and medial gastrocnemius. Together, these included areas innervated by nerve roots from L4 through S2. Bilateral mapping needle EMG of the paraspinal muscles was done using the MiniPM technique.Masked opinions from the physiatrist, neurosurgeon, and neuroradiologist were used to identify a consensus “gold-standard” subgroup of 55 subjects for whom there was unanimous agreement. Follow-up of this subgroup was done at more than one year. Haig reported that paraspinal-mapping EMG had 100% specificity and 30% sensitivity for detecting stenosis compared with either the back-pain or asymptomatic control groups (p<0.04). Combining limb and paraspinal-fibrillation scores had a sensitivity of 47.8% and specificity of 84.5% (p=0.008). EMG identifies neuromuscular diseases that pose as stenosisEMG was also able to detect some neuromuscular diseases that masquerade as stenosis and cannot be diagnosed using imaging alone. In the 55-patient "gold-standard" subgroup, EMG detected six cases of polyneuropathy and one case of myopathy. Five of these were in patients originally judged to have stenosis, and two were in patients from the asymptomatic control group. These neuromuscular diseases included one Charcot-Marie-Tooth disease, one statin myopathy, three patients who had been diagnosed with diabetes and were presumed to be diabetic neuropathies, and two nonspecific polyneuropathies. "Some of these are reversible, but none of these are amenable to operation," Haig said. "Low back pain doesn't require a 'routine workup' aside from the history and physical examination, but when the pain doesn't go away for a month or when there are medical red flags, then EMG is an important part of the diagnostic workup. EMG is more properly called an electrodiagnostic consultation, since the whole process is quite complex and requires substantial specialty training," Haig said. If EMG is so good, why is MRI in all the guidelines?Haig also told rheumawire that subsequent work from this study presented in abstract form at the International Society for Physical and Rehabilitation Medicine (Sao Paulo, May 2005) and now in revision for journal publication showed that MRI could not successfully differentiate people whom clinicians thought had spinal stenosis from asymptomatic older people. This raises the question of why EMG went largely unmentioned in previous comprehensive reviews of spinal-stenosis diagnosis and treatment. Haig attributes this to "ignorance hidden behind knowledge." "Surgeons, who typically know nothing of EMG, often lead or predominate in [writing] spine guidelines. When they want expertise in diagnostics, they call in the radiologists. I have no idea what the guideline reviewers do with the 100 or so articles on EMG and spinal diagnostics. The EMG articles have not all been of great quality, but they're at least as good as the surgical literature," Haig said. He also points out that there are "thousands and thousands" of electromyographers in every community hospital in the country. "This article meets most people's criteria for a stringent scientific study, so I hope EMG will get in the back door somehow. It can prevent bad things from happening to people with back pain." (Source: [1] Haig AJ, Tong HC, Yamakawa KSJ, et al. The sensitivity and specificity of electrodiagnostic testing for the clinical syndrome of lumbar spinal stenosis. Spine 2005; 30:2667-2676: Joint and Bone: Rheumawire: December 2005.)