Diabetic Sensorimotor Autonomic Polyneuropathy

Diabetic neuropathies are by far the most prevalent peripheral neuropathies encountered in clinical practice. Among all diabetic neuropathies, the mixed sensory-motor-autonomic peripheral polyneuropathy is by far the most common, and is usually related to the duration and severity of hyperglycemia. However, this form may occasionally be the presenting symptom of occult diabetes mellitus. The exact incidence of this diabetic polyneuropathy is not known; this is, in part, because of its diverse clinical presentations and different measurements used to define the presence or absence of neuropathy (Table C7-2). The reported incidence of diabetic neuropathy in general is extremely variable, ranging from 5 to 50+. In a large cohort study followed for 25 years, it was estimated that 8+ of diabetics have neuropathy at the time of diagnosis, and neuropathy develops in 50+ of patients within 25 years of diagnosis (see Pirart 1978).

Table C7-2 Available Measures That May Be Used to Assess for Diabetic Peripheral Polyneuropathy

The clinical manifestations of mixed diabetic polyneuropathy are due to an axonopathy and follow a length-dependent pattern, with the longest axons involved first distally. The sensory manifestations often begin in the toes and progress slowly cephalad to the distal legs. They generally do reach the fingertips or hands when the lower limb symptoms are at around the level of the knees. Sensory symptoms due to loss of small fibers usually appear first, as paresthesias and neuropathic pain (deep aching or throbbing and superficial burning or stabbing). Pain and temperature sensation are usually blunted on examination. Impaired leg proprioception and vibration sense and diminution of ankle jerks, manifestations of large fiber loss, occur later in the course of the illness.

Motor fiber involvement in mixed diabetic polyneuropathy is minimal in the early stages of the disease. This is typically confined to weakness and atrophy of the intrinsic foot muscles and weakness of toe flexors and extensors. However, when weakness worsens, it follows also a distal to proximal gradient resulting in progressive foot weakness sometimes leading to bilateral flail foot, and hand weakness and atrophy occasionally causing bilateral clawed hands.

Manifestations of autonomic dysfunction become increasingly important as the neuropathy progresses, and occasionally they dominate the clinical picture. This includes impotence due to impaired dysfunction, abdominal pain, diarrhea, and constipation due to impaired gastrointestinal motility, distal anhidrosis due to sudomotor dysfunction, and orthostatic intolerance due to vasomotor denervation. More advanced dysautonomia may manifest as recurrent syncope due to orthostatic hypotension or neurogenic bladder resulting in reduced sensation of bladder fullness and incomplete emptying. Loss of hypoglycemia warnings results from the combination of adrenal gland denervation (leading to blunted catecholamines excretion), sudomotor denervation (leading to loss of sweating), and cardiac denervation (leading to the loss of reactive tachycardia). Finally, acrodystrophic changes of skin, nails, and joints may dominate due to small fiber sensory loss, with accompanying foot ulcerations and neuropathic arthropathy (Charcot joint) of the ankle and foot.

Diabetic Amyotrophy

Diabetic amyotrophy is a much less common neuropathy than the chronic mixed sensorimotor diabetic polyneuropathy. The term was first coined in 1953 by Garland and Taverner. They first called this syndrome “diabetic myelopathy,” because they presumed that the pathology was in the spinal cord, particularly in the anterior horn cell column. This disorder has been surrounded by controversy primarily because of lack of understanding of the exact site and nature of the pathology, which has been attributed to lesions of the anterior horn cells, lumbar roots, lumbar plexus, and femoral nerve. Authors have thus ascribed many terminologies to this disorder, based on their own theory of the nature of the illness (Table C7-3). Although many have suggested that the name “diabetic amyotrophy” be abandoned because of its ambiguity, the term continues to be the most commonly used for this disorder in neurologic practice. Another popular designation is “subacute diabetic proximal neuropathy.”

Table C7-3 Common Synonyms of Diabetic Amyotrophy

Adapted with revisions from Wilbourn AJ. The diabetic neuropathies. In: Brown WF, Bolton CF, eds. Clinical electromyography, 2nd ed. Boston, MA: Butterworth-Heinemann, 1993.

Diabetic amyotrophy affects primarily the L2, L3, and L4 roots/plexus. This may be a unilateral, bilaterally asymmetrical, or bilaterally symmetrical condition. Sometimes, additional adjoining roots become involved, such as the lower thoracic roots (leading to diabetic thoracic radiculopathy), or the L5 and S1 roots (resulting in so-called diabetic footdrop).

The typical patient is a type 2 diabetic adult, usually older than 50 years of age, in whom a subacute unilateral proximal leg weakness develops that evolves over several weeks to months. Thigh pain is extremely common; it usually is deep and worse at night, but it can be burning and severe. At times, it involves the buttock and back. Weakness invariably involves the quadriceps, iliopsoas, and thigh adductor muscles but may extend into the tibialis anterior, glutei, hamstrings, and, rarely, the gastrocnemius. Knee jerk is depressed or absent. Not infrequently, there is clinical or only electrophysiologic evidence of contralateral involvement, usually milder in degree. Also, the contralateral leg may become affected later, while the initial ipsilateral weakness is improving.

The site of the lesion and the exact pathophysiology of diabetic amyotrophy are not well known, partially because of the lack of adequate pathologic studies. An inflammatory vasculopathy (vasculitis or perivasculitis) causing ischemic nerve infarction is the most popular theory and is supported by a single autopsy (see Raff et al. 1968) and several nerve biopsy series injury (see Said et al. 1994, 1997). Some observers distinguish between diabetic patients with proximal neuropathy in whom an asymmetrical neuropathy rapidly develops and those with more gradually progressive symmetrical neuropathy. These authors have proposed that the former syndrome is ischemic (vascular) in nature, and the latter metabolic. In practice, many patients fit into a spectrum between these two extremes, making the distinction difficult and two separate mechanisms unlikely.

The following are frequently asked questions regarding diabetic amyotrophy:

The primary differential diagnosis of diabetic amyotrophy is an L2–L4 radiculopathy due compressive spinal lesion. Night pain and allodynia favor diabetic amyotrophy. Imaging of the spine is often necessary for confirmation. When quadriceps weakness is severe, spinal compressive disease must include at least two lumbar roots. Other considerations involve other causes of lumbar plexopathy, femoral neuropathy, and motor neuron disease.

Diabetic Sensorimotor Autonomic Polyneuropathy

Electrodiagnostic (EDX) studies are valuable in confirming the presence of chronic axon loss peripheral polyneuropathy. During the early stages of the disorder and in a small percentage of patients whose manifestations are restricted to the toes or small fiber symptomatology, the NCS and needle EMG examination, which assess only the large myelinated nerve fibers, may be normal. In these situations, other modalities such as quantitative sudomotor axon reflex test, quantitative sensory testing, or skin biopsy may be necessary to show involvement of small unmyelinated fibers. When large fibers undergo axonal degeneration, the EDX abnormalities are initially found only in the lower extremities. These typically consist of one or more of these NCS abnormalities: absent H reflexes, low-amplitude or absent sural and superficial peroneal SNAPs, low-amplitude tibial and peroneal CMAPs, and mild slowing of peroneal and tibial motor distal latencies and conduction velocities (Figure C7-1). The needle EMG often shows long-duration and high-amplitude MUAPs with or without fibrillation potentials in the intrinsic foot muscles. With more advanced disease, the neurogenic MUAP changes worsen in the leg and ascend that abnormalities are found in the upper limbs. Initially, this usually presents as reduction of the median, ulnar, and radial SNAP amplitudes with mild slowing, low, or borderline median and ulnar CMAPs with mild sensory and motor conduction slowing, with long-duration and high-amplitude MUAPs with or without fibrillation potentials in the intrinsic hand muscles. In severe polyneuropathy there is often complete absence of all routine sensory and motor conduction studies in the legs and hands, with long-duration, high-amplitude and rapidly recruited MUAPs with or without fibrillation potentials in all the leg and arm muscles, which are worse distally.

Figure C7-1 Peroneal motor nerve conduction study recording extensor digitorum brevis in a 50-year-old patient with a 7-year history of type 2 diabetes mellitus and chronic distal peripheral polyneuropathy (B), compared to an age-matched control (A). Note that the distal CMAP amplitude is significantly lower in the diabetic patient than the control (3.8 mV versus 10.0 mV) while the distal latencies and conduction velocities are only marginally slowed, consistent with an axonopathy (axon-loss peripheral polyneuropathy).

The EDX features of diabetic sensorimotor peripheral polyneuropathy are characteristic of a primarily axon loss polyneuropathy. However, the abnormalities are not specific for diabetes mellitus and are encountered with a wide variety of other metabolic or toxic etiologies. The EDX are most useful when the cause of the neuropathy is unclear, and the EDX studies become essential in looking for demyelinating features, which would suggest another diagnosis, such as an acquired or familial demyelinating polyneuropathies. The EDX studies are also very useful in diabetic patients with disproportionately upper extremity symptoms in order to exclude entrapment neuropathies such as the carpal tunnel syndromes.

Diabetic Amyotrophy

Routine sensory and motor nerve conduction studies are abnormal, particularly in the legs, in a majority of patients with diabetic amyotrophy because there usually is a concomitant diabetic sensorimotor peripheral polyneuropathy. Based on electrophysiologic criteria, two-thirds of patients with diabetic amyotrophy have an associated distal peripheral polyneuropathy. In addition to the peroneal and tibial motor NCS, the femoral CMAP should be obtained in all patients with suspected diabetic amyotrophy. Although this study adds little to the diagnosis, it is very useful in assessing the degree of axonal loss. In diabetic amyotrophy, it is frequently low in amplitude, unilaterally or bilaterally, which is consistent with the axonal nature of the lesion. Slowing is, however, minimal. The saphenous SNAPs also should be done bilaterally, although they often are unelicitable in both legs in these patients because of age, obesity, or concomitant distal sensorimotor peripheral polyneuropathy.

Needle EMG findings in diabetic amyotrophy reflects the consequences of axonal loss, but are heterogeneous in distribution and chronicity. Most commonly, the needle examination reveals fibrillation potentials, decreased recruitment, and large and polyphasic MUAPs in an L2 through L4 roots distribution. The muscles that are usually most involved include the quadriceps (L2–L3–L4), the thigh adductors (L2–L3–L4), and the iliacus (L2–L3), with less prominent changes noted in the tibialis anterior (L4–L5). In patients with an overlapping distal diabetic chronic polyneuropathy, there often is distal denervation also, particularly in the abductor hallucis, extensor digitorum brevis, extensor hallucis, and flexor digitorum longus. Thus, when needle EMG is being performed on patients with diabetic amyotrophy, starting distally in the foot and proceeding proximally, fibrillation potentials and loss of MUAPs are often noted in the foot muscles and the anterior thigh muscles but not in the leg muscles. This creates a typical “skip region,” which includes the medial gastrocnemius and sometimes the tibialis anterior also. This region is surrounded by the abnormal muscles; the denervation in the intrinsic foot and distal leg muscles is caused by the distal polyneuropathy, while the denervation in the quadriceps, thigh adductors, and iliacus proximally is caused by the diabetic amyotrophy.

In some patients, the subacute denervation of diabetic amyotrophy extends caudally into the L5 root. In these patients, fibrillation potentials with loss of MUAPs are also prominent in the tibialis anterior, the extensor hallucis (resulting in so-called diabetic footdrop), the tibialis posterior, the flexor digitorum longus, and, occasionally, the hamstrings and glutei. In severe cases, the denervation may become so diffuse in the lower extremity that few muscles escape the condition. Among these, the medial head of the gastrocnemius (S1–S2) is the most likely to be free of fibrillation potentials; however, it might harbor chronic neurogenic changes when an overlapping chronic distal diabetic polyneuropathy coexists.

Regardless of the clinical picture (unilateral or bilateral), it is common to find abnormalities on needle EMG in the contralateral muscles of many patients with diabetic amyotrophy. Even in patients with symmetrical weakness at the time of the EMG examination, the neurogenic changes on needle EMG are often different, suggesting a different time course of the processes in both legs. Typical needle EMG findings include profuse fibrillation with highly polyphasic MUAPs in one thigh, consistent with a subacute process, and rare fibrillations with many polyphasic MUAPs of increased duration and amplitudes in the contralateral thigh, consistent with a more chronic disorder.

Based on the aforementioned findings, an extensive needle EMG is often required in patients with suspected diabetic amyotrophy. Sampling the quadriceps (preferably at least two heads), the thigh adductors, the iliacus, the tibialis anterior, and the lumbar paraspinal muscles is essential. Because the disorder is often bilateral, testing the same muscles on the contralateral side in symptomatic limbs or, at least, the quadriceps and iliacus in asymptomatic limbs is recommended. Finally, sampling more distal muscles is essential when there is possible concomitant distal polyneuropathy.

It is often necessary to perform imaging studies, such as MRI, on the lumbar spine to exclude a compressive lesion within the lumbar intraspinal canal, because differentiating diabetic amyotrophy from L2–L4 intraspinal lesions is difficult, even for the seasoned electromyographer, for several reasons:

Until recently, it was widely accepted that isolated femoral mononeuropathy is a complication of diabetes mellitus. However, it is now clear that this is a misnomer. It is likely that most reported cases, published more than 30 years ago, actually involved mislabeled patients with diabetic amyotrophy; this occurred because many physicians and electromyographers did not assess other muscles thoroughly, in particular the thigh adductors. Although the brunt of weakness in many patients with diabetic amyotrophy often falls on the quadriceps muscle mimicking selective femoral nerve lesions, careful clinical and needle EMG examinations reveal more widespread involvement of thigh adductors and sometimes foot dorsiflexors, muscles not innervated by the femoral nerve. Despite the current knowledge, the term diabetic femoral neuropathy, unfortunately, has not completely vanished (see Coppack and Watkins 1991).