EMG

The most reliable of all available electrodiagnostic methods for detecting radiculopathy is the needle EMG. Although neuroimaging studies are usually diagnostic in the more common radiculopathies caused by structural lesions, they may be unremarkable in radiculopathies caused by infection, infiltration, demyelination, or infarction.

The diagnosis of radiculopathy through EMG is based on demonstrating evidence of ongoing denervation or chronic reinnervation in at least 2 muscles innervated by the same nerve root, but innervated by different peripheral nerves. EMG abnormalities must be in a myotomal distribution. Clinically weak muscles should be preferentially examined during EMG. Ongoing denervation is manifested by the presence of fibrillations, positive sharp waves, and reduced recruitment. Polyphasic motor unit action potentials (MUAPs) may be seen during the period of early reinnervation. Chronic reinnervation is manifested by the presence of MUAPs with long durations and large amplitudes. Not all muscles innervated by the involved root need to be abnormal on EMG. However, it is important to demonstrate that the muscles innervated by unaffected adjacent nerve roots do not have ongoing denervation or chronic reinnervation.

The optimal number of muscles that should be examined when screening for a lumbar radiculopathy is at least 5 lower extremity muscles plus a paraspinal muscle (for a total of 6 muscles). If the paraspinals cannot be reliably tested, then at least 8 nonparaspinal muscles must be examined. The muscles examined should be representative of the key myotomes on the extremity. Several investigators have published tables of muscles that show a high yield for detecting radiculopathy at specific root levels. If a particular root is already suspect based on clinical and radiographic data, then additional muscles innervated by that root but by different peripheral nerves should be examined. While the distribution of EMG abnormalities is expected to be in a myotomal pattern, the difficulty lies in the fact that most muscles are innervated by more than one root. Thus, it is sometimes impossible to determine electrophysiologically which of 2 contiguous roots is involved. For example, the L2, L3, and L4 roots overlap extensively in their muscular innervations. Furthermore, EMG cannot be used to localize the exact anatomic site of a pathologic condition because a single root may be compressed by a disk or osteophyte at more than one level.

For abnormalities to be seen on EMG, the underlying pathologic condition must involve motor axon loss. A purely sensory radiculopathy would have a normal EMG. Likewise, a purely demyelinating radiculopathy without axonal loss (which is rare) would also have a normal EMG. On the other hand, a demyelinating radiculopathy with conduction block would present clinically with weakness and the EMG would demonstrate decreased recruitment.

The paraspinal muscles should be examined during the radiculopathy workup. The presence of ongoing paraspinal denervation, in the form of fibrillations and positive sharp waves, is an important localizing finding because it indicates that the site of nerve root abnormality is proximal to the dorsal ramus takeoff, and thus situates the lesion at the root or the anterior horn cell level. Unfortunately, findings of paraspinal denervation do not help determine the precise segmental level of the lesion because there is much overlap in the innervation of these back muscles. Furthermore, the presence of paraspinal denervation must be interpreted carefully because this is also seen in patients with conditions other than radiculopathy, such as motor neuron diseases and after low back surgeries and injections. Abnormal spontaneous activity in the paraspinal muscles may occur up to 4 days after a lumbar puncture or myelography. Paraspinal EMG may be abnormal more than 3 years after lumbosacral surgery. Mild degrees of paraspinal denervation may be seen in normal asymptomatic individuals, likely reflecting root injury caused by normal age-related degeneration of the spine.

To further add to the confusion, about half of radiculopathies have normal paraspinal EMGs. One explanation is the possible fascicular sparing of fibers to the dorsal rami. Moreover, the absence of paraspinal denervation cannot totally exclude a radiculopathy, as the paraspinal muscles may have been spared in incomplete root lesions or the paraspinals may have already reinnervated in cases of chronic radiculopathies. There are also several technical issues that may hinder the detection of paraspinal denervation, such as difficulty of the patient to either relax or activate the back muscles during EMG, or simply sampling limitations.

Electrodiagnostics is a time-sensitive test; therefore, the timing of electrodiagnostics affects the results and their interpretation because of the natural history of radiculopathy. Soon after an acute nerve root injury, no abnormalities may be detected by electrodiagnostics except perhaps for reduced recruitment and/or abnormal late responses (H-reflex and F-wave). It may take 7 to 10 days for fibrillation potentials to appear in the paraspinal muscles and 3 to 6 weeks to appear in the limb muscles, with abnormal findings progressing in a proximal to distal manner. It may take 6 to 26 weeks for MUAPs to display variable configurations or polyphasia, indicating the start of reinnervation. Chronic reinnervation occurs months after injury and is manifested by the presence of MUAPs with long duration and large amplitude (so-called neuropathic MUAPs.) Reinnervation, like denervation, progresses in a proximal to distal manner. Reinnervation changes are first seen in the paraspinal and proximal limb muscles by around 3 months postinjury, and then are later seen in the distal muscles by around 6 months postinjury. Electrodiagnostics is better suited to diagnosing subacute radiculopathies (3 weeks to 3 months from onset) rather than acute or chronic radiculopathies. Thus, electrodiagnostics should ideally be performed at least 3 weeks from the onset of symptoms (not earlier) to allow time for most limb muscles to develop signs of axon loss (fibrillation potentials). Otherwise, electrodiagnostics performed too early may yield inconclusive results and may require subjecting the patient to another round of tests a few weeks later.

EMG by itself can already meet most of the secondary objectives of electrodiagnostics. EMG can determine the nerve roots involved by demonstrating the myotomal distribution of abnormalities. EMG abnormalities, such as abnormal spontaneous activity (fibrillations and positive sharp waves) and neuropathic MUAPs, substantiate the existence of an underlying axonal abnormality. The severity of the radiculopathy corresponds to the gradation of abnormal spontaneous activity present and the amount of MUAP dropout. The chronicity of the radiculopathy can be roughly estimated based on the proximal-distal distribution of abnormal EMG findings along an affected myotome, and based on the size of fibrillations (which are larger in newer lesions and smaller in older ones).

NCS

NCS are usually normal in radiculopathies. Nevertheless, it is necessary to perform NCS during a radiculopathy workup to exclude other conditions that may mimic radiculopathy, such as a plexopathy, polyneuropathy, or entrapment neuropathy. For example, when assessing for L5 radiculopathy it is important to rule out a peroneal (fibular) neuropathy at the fibular neck because both conditions may present with similar signs and symptoms (foot drop and numbness on the dorsum of the foot). The following NCS are commonly performed during a lumbosacral radiculopathy workup: sural and superficial peroneal (fibular) sensory studies, peroneal (fibular) and tibial motor studies (often with F-waves), and tibial H-reflexes.