Clinical manifestations

Presentation and course

An important fundamental clinical point is that even though sensorimotor polyneuropathy is the most frequent of the diabetic neuropathies, many patients have more than one of the conditions or types of neuropathy listed in Table 1, either sequentially or concomitantly. Both type 1 and type 2 diabetics are at risk for these neuropathies.

Distal symmetric sensorimotor polyneuropathy. This is the most common form of diabetic neuropathies. Many physicians inadequately assume that the term diabetic neuropathy is synonymous with distal symmetrical polyneuropathy because the latter constitutes about three fourths of all diabetic neuropathies. Symptoms are mainly sensory and insidious in onset. The prominent symptom is often reduced sensation, but there may also be paresthesias, tingling, burning, and neuropathic pain. Indeed, these "negative" and "positive" sensory symptoms may coexist. Symptoms begin distally in the toes and the feet and gradually extend proximally. Later, the fingers and hands may become affected, again with proximal spread. When extensive, the anterior abdominal wall is involved, and the sensory loss gradually spreads laterally around the trunk. Motor involvement is less frequent than sensory. However, when severe, this neuropathy causes weakness of distal leg muscles and later of the intrinsic hand muscles. Unsteady walking due to sensory ataxia is the result of large fiber sensory loss with or without concomitant weakness. Although this chronic neuropathy is related to the duration and severity of hyperglycemia, it can occasionally be the presenting symptom of occult diabetes mellitus, and significant sensorimotor and autonomic abnormalities may occur in patients with mild degrees of hyperglycemia. Patients with chronic and severe polyneuropathy have increased risk of foot ulceration, which contributes to lower extremity amputations in diabetic patients. This is due in part to a combination of lack of protective sensation, abnormal sweating, and poor wound healing.

Episodic or transient presentation. Three uncommon and unusual subacute sensory neuropathy syndromes are characterized by their eventual recovery:

The clinical examination for polyneuropathy focuses on demonstrating impaired distal sensation distally, hypo- or areflexia, and weakness. Because the sensory fiber involvement may be restricted (see below), it is important to test for small fiber sensory loss using pinprick and temperature testing as well as for large fiber loss--touch, vibration, reflex loss, and weakness. The use of monofilaments to assess touch-pressure sensation has become popular (140).

Diabetic autonomic neuropathy. Autonomic nerve fibers are invariably involved in chronic sensorimotor polyneuropathy. This is frequently subclinical in the early stages of the polyneuropathy, although it may be detected using sensitive methods to measure and quantitate autonomic function. When symptomatic, this may result in impaired sweating and some cutaneous vasomotor changes. However, the autonomic nervous system may become widely involved and dominate the clinical picture. In most patients the symptoms are not severe, but some have devastating diabetic autonomic neuropathy (139). The neuropathy may affect all or selected organs or systems innervated by the autonomic nervous system. In autonomic neuropathies, cardiovascular, gastrointestinal, and urogenital systems may be affected and sudomotor function impaired. Thus, one or more of the following may develop: gastroparesis, diarrhea, constipation, orthostatic hypotension, bladder dysfunction, and erectile dysfunction. About 40% of diabetic men develop erectile dysfunction; this may occur in the absence of, or in association with, other manifestations of diabetic autonomic neuropathy.

The clinical examination of the autonomic nervous system is limited. Cardiovascular autonomic testing is most simply performed by evaluating the heart rate. A fixed heart rate on deep breathing has the greatest specificity for cardiac parasympathetic dysfunction (approximately 80%). A resting tachycardia, or when the patient goes from lying to standing, also indicates vagal parasympathetic dysfunction. The simple bedside measurement of lying-standing blood pressure change is an important test for sympathetic vasoconstrictor dysfunction. Dry feet indicate a failure of distal sweating. Reduced lacrimation can be detected using Schirmer strips.

Diabetics with cardiac autonomic neuropathy have significantly increased overall mortality (141). Hence, autonomic assessment may be used for cardiovascular risk stratification in diabetic patients, and as indicator for more aggressive pharmacologic therapies and lifestyle management of comorbid conditions.

Diabetic small fiber neuropathy. Damage to the small sensory nerve fibers is one of the earliest manifestations of diabetic neuropathy. This often starts with episodic pain in toes that slowly spreads to the feet and sometimes above ankles and fingers. The type of neuropathic pain associated with small fiber neuropathy is variable and often mixed, including burning, electrical shocks, tingling, pins and needles, painful cold, uncomfortable numbness, and frequent itching. On examination, there is often touch and pin hypoesthesia, tactile hyperesthesia or allodynia, or both. These two findings when combined have excellent sensitivity (83%) and specificity (90%) for the diagnosis of neuropathic pain and diabetic small fiber neuropathy (124). There is evidence suggesting parallel damage to small and large nerve fibers (152).

Cranial neuropathies. Cranial nerve palsies are rare acute complications of diabetes. The most common cranial neuropathy is third nerve palsy (60). This typically affects patients over the age of 50 years and is rare in children. Only one fourth of patients have an associated diabetic peripheral neuropathy (60). It is usually abrupt and associated with pain. Pupillary function is usually spared. Pupillary sparing has been attributed to ischemia occurring centrally within the third nerve, preserving the peripherally located parasympathetic pupilloconstrictor fibers. This contrasts with compressive lesions of the oculomotor nerve, such as aneurysm of the posterior communicating artery, in which the pupillary fibers usually are affected early. Sixth nerve palsies also occur in diabetics. It is unclear whether seventh nerve palsies occur more frequently in diabetics.

Truncal neuropathy. Truncal or thoracic neuropathy usually presents subacutely with painful paresthesias in variable size patches unilaterally or bilaterally in the trunk (125). This usually involves the T4 through T12 spinal nerve roots individually or over several contiguous nerves. Associated involvement of motor nerve fibers can lead to bulging of the abdominal wall in the paresthetic areas, often referred to as pseudohernia and best appreciated when the patient is standing. Clinical examination usually reveals a patch of sensory abnormality in the region of the symptoms. Abnormalities consist of varying degrees of sensory blunting that can be subtle or hyperalgesia or allodynia.

Focal limb neuropathies. Many studies have shown that peripheral nerve compressive lesions are more frequent in diabetics than in nondiabetics. They add to the disability already imposed by the polyneuropathy that is almost always present. Diabetes mellitus has been long considered a risk factor for carpal tunnel syndrome, but this remains controversial (56; 64). Ulnar neuropathies in diabetics are often insidious and are mainly motor with limited sensory symptoms and signs. They are often due to nerve compression at the elbow, though patients with ulnar nerve lesions within the forearm segment showing partial conduction block or abnormal temporal dispersion on nerve conduction studies are increasingly reported in diabetics. Such focal neuropathies can easily go undetected because the symptoms are attributed to polyneuropathy. When sensory or motor symptoms are more prominent in the hands than feet, carpal tunnel syndromes or ulnar neuropathies should be suspected and excluded. Meralgia paresthetica (mononeuropathy of the lateral femoral cutaneous nerve) is associated with diabetes mellitus irrespective of obesity and advanced age (101). The incidence rate in the diabetic population is more than 7 times the general population. Patients with meralgia paresthetica are also at-risk for developing diabetes, and counseling these patients in prevention of diabetes is warranted.

Diabetic proximal (proximal diabetic neuropathy, diabetic radiculoplexus neuropathy, diabetic amyotrophy). This syndrome "goes by a bewildering variety of names" (38). These include diabetic amyotrophy, diabetic lumbosacral plexopathy, diabetic polyradiculopathy, proximal diabetic neuropathy, ischemic mononeuropathy multiplex associated with diabetes mellitus, Bruns-Garland syndrome, and diabetic lumbosacral radiculoplexus neuropathy (12; 41). It is also misleadingly referred to as “diabetic femoral neuropathy.”

This condition may develop in longstanding diabetics during periods of poor metabolic control and weight loss, but it can also occur in mild and well controlled diabetics or be the presenting feature of diabetes. Diabetic lumbosacral radiculoplexopathy may be treatment-induced, manifesting during aggressive treatment of diabetes (121; 142). The clinical features are variable, and the onset may be gradual or sudden. A prominent feature is pain that is often severe and located in the back, hips, and thighs; occasionally, pain is mild or even absent. The patient may have the systemic symptoms of anorexia, malaise, and weight loss.

On examination, the usual motor features are unilateral wasting and weakness of the proximal legs and hip girdle, but distal muscles may also be affected. The quadriceps muscle group is often most affected. The weakness varies from mild to so severe as to render the patient wheelchair-bound. A diagnosis of femoral neuropathy is often made, but detailed clinical and electrophysiological tests will usually show more extensive involvement. These patients may be misdiagnosed as having lumbosacral radiculopathies, leading to unnecessary surgeries before the neurologist identifies the correct diagnosis. Proximal sensory involvement is variable and often minor compared with the motor abnormalities. All these symptoms and signs are usually, but not always, asymmetrical. Patients frequently have coexisting sensorimotor polyneuropathy.

A similar disorder without pain though sharing many of the other features of the painful syndrome (weight loss, proximal weakness, subacute onset) has been also described. This variant, however, tends to have more bilateral and symmetrical onset, is more slowly progressive, and often results in significant disability (32; 55).

Patients with a similar clinical entity sharing many of the clinical and pathological features of diabetic lumbosacral radiculoplexus neuropathy but affecting the upper extremity are rarely described. These patients have a disorder that mimics idiopathic neuralgic amyotrophy (brachial neuritis) with acute pain followed by weakness. The term “diabetic cervical radiculoplexus neuropathy” has been proposed (93).

Superimposed chronic inflammatory demyelinating polyneuropathy (CIDP). Although a study suggested otherwise (81), it is possible that a chronic inflammatory demyelinating polyneuropathy-like syndrome may occur more frequently in diabetics than in nondiabetics (79; 127; 71; 59; 35; 120). Features that distinguish chronic inflammatory demyelinating polyneuropathy from the far more frequent chronic diabetic sensorimotor polyneuropathy include rapidly progressive neuropathy, the presence of oligoclonal bands in cerebrospinal fluid, demyelinating features on nerve conduction studies, proximal nerve and brachial plexus enlargements on ultrasound (128), and a good response to a therapeutic trial of immunomodulatory treatments (intravenous immunoglobulin or corticosteroids). Unfortunately, high CSF protein also may be found in patients with diabetes without chronic inflammatory demyelinating polyneuropathy (77). Although proportion of treatment responders was comparable in patients with and without diabetes (70% vs. 74.9%, respectively), a study showed a class I evidence that IVIg does not reduce disability in patients with chronic inflammatory demyelinating polyneuropathy and diabetes (19; 113). Making a diagnosis of true chronic inflammatory demyelinating polyneuropathy in diabetic patients remains challenging because demyelinating features may be present on electrophysiology in patients with diabetes. Hence, these electrophysiologic findings should not be used as the sole criteria (absent clinical findings) for the diagnosis of an immune-mediated neuropathy such as CIDP.

Impaired glucose tolerance (prediabetes) and neuropathy. Increasingly, prediabetes has been regarded as a cause of painful peripheral neuropathy. Patients with mild, predominantly sensory neuropathy, but with normal fasting blood glucose, may have impaired glucose tolerance on oral glucose tolerance testing. The prevalence of neuropathy in patients with prediabetes may be higher than in nondiabetics, though this is still a subject for debate (25; 83). Given commodities in prediabetes, contributions from the metabolic syndrome as the cause of neuropathy should be also considered (73).

The diagnosis of impaired glucose tolerance (prediabetes) remains challenging and often is under dispute by both physicians and patients. The diagnostic sensitivity of hemoglobin A1c for prediabetes is unclear, although the specificity is very high. In at-risk patients with prediabetes and normal hemoglobin A1c, oral glucose tolerance test is recommended (25). The American Diabetes Association recommends that two tests (fasting glucose, HbA1c measurement, 2-hour plasma glucose level after an oral glucose tolerance test) obtained at two separate visits should be abnormal for confirmation of the diagnosis of diabetes (06). Elevated levels of fasting glucose and HbA1c from a single baseline blood sample was shown to have moderate sensitivity (55%) but high specificity to identify patients who will develop frank diabetes in the next 5 years of follow-up (98%), with specificity increasing to 99.6% at 15 years (119).

Prognosis and complications

Diabetic sensorimotor polyneuropathy usually worsens slowly over the years. About half of all diabetic neuropathies are asymptomatic and remain only mildly symptomatic in most patients (108). Others may be severely troubled by one or more of the following complications: severe neuropathic pain, numbness and other sensory symptoms in the feet and hands, distal limb weakness late in the course of disease, and even falls due to proprioceptive loss. Falls are frequent in older diabetic patients, particularly with lower extremity weakness or vibration loss (65). These patients should be counseled and undergo gait evaluation, walking aids, and physical therapy. Polyneuropathy is a risk factor for foot ulcers, Charcot joints, and amputation.

The symptoms of hyperglycemic neuropathy resolve rapidly with the establishment of euglycemia (61). Acute painful sensory neuropathy usually improves after several weeks or months. Truncal neuropathy will also usually resolve, although this may take many months. Some patients have repeated episodes, essentially having chronic pain.

Diabetic autonomic neuropathy is generally considered to be a progressive disorder and has long been suspected as being a risk factor for renal failure, myocardial infarction, and sudden unexpected death (117; 144). Impaired cardiovascular autonomic function as measured by heart rate variability is associated (ie, relative risk is doubled) with an increased risk of silent myocardial ischemia and mortality (139). One study showed that sudden cardiac death in diabetics was correlated with atherosclerotic heart disease and nephropathy and to a lesser degree with diabetic autonomic neuropathy; the latter is unlikely to be the primary cause of cardiac death (90).

In proximal diabetic neuropathy, the weakness initially worsens over weeks or sometimes months, although the pain resolves earlier. However, this weakness eventually improves spontaneously to some extent after 3 to 18 months. Mild weakness, discomfort, and stiffness often persist for years, and only 40% of patients make a full recovery (36; 38). Relapses on the same or other side occasionally occur.

Both the third and sixth cranial neuropathies almost always recover completely or nearly so in a period of a few weeks to a few months. Whether the prognosis of seventh nerve palsy (Bell palsy) is different in diabetics compared to the general population remains unresolved.

Carpal tunnel syndrome in diabetics should be managed exactly as in nondiabetics, ie, wrist splints or decompression (02). Improvement, at least back to the baseline polyneuropathy symptoms, if present in the hands, is to be expected following decompression. No data exist regarding the results of surgical interventions for ulnar and other focal neuropathies in diabetics.

The prognosis for the rare case of diabetes associated with definite clinical evidence of chronic inflammatory demyelinating polyneuropathy (136) is generally good when it is treated with immunomodulatory therapy. It is hoped that the treatment of impaired glucose tolerance will improve or stabilize the neuropathy, but addressing causes of the metabolic syndrome are also important (73).

Clinical vignette

A 35-year-old man had type 1 diabetes since the age of 13 years. At the time of presentation, he had retinopathy for 5 years and nephropathy for 4 years. Sensorimotor polyneuropathy was recognized at the age of 30, and he developed a Charcot joint of the left ankle at the age of 34.

Autonomic dysfunction began at 25 years of age with erectile dysfunction, which was managed with an implant. He later developed intermittent diarrhea but was occasionally constipated. Bladder function remained normal. Orthostatic hypotension on standing up from bed was troublesome, so he sat before standing. Syncopal episodes were increasing. Sweating was also abnormal (from the neck down to the mid-thighs but dry everywhere else).

Examination showed moderate distal wasting and weakness in the limbs as well as complete areflexia. Light touch and pinprick sensations were diminished in the tips of the fingers as well as in the feet to the mid-calf. Vibration and proprioception sensations were absent in the toes. Blood pressure was 160/90 mm Hg lying and 100/60 mm Hg standing. His pulse was 90 per minute in both positions. When sitting in a warm examination room, his hands and feet were completely dry, but his trunk was excessively sweaty, and his shirt became completely wet.

Improved diabetes control was attempted, but the patient was not compliant. The erectile dysfunction was managed with an implant, diarrhea with diphenoxylate hydrochloride, and the orthostatic hypotension with fludrocortisone.

Comment. This young type 1 diabetic had moderate to severe sensorimotor polyneuropathy, leading to foot complications. Erectile dysfunction developed without, at first, other manifestations of diabetic autonomic neuropathy, but chronic diarrhea, orthostatic hypotension, fixed tachycardia (lack of vagal innervation to the heart), and sweating abnormalities developed later. Patients with these neurologic syndromes frequently have other diabetic complications such as retinopathy and nephropathy.

Biological basis

Etiology and pathogenesis

Pathogenesis. In spite of massive research efforts, the pathogenesis of diabetic neuropathies remains incompletely understood. Several different and sometimes overlapping mechanisms contribute to nerve damage. The different types of neuropathy probably have different pathogenic mechanisms.

In chronic sensorimotor polyneuropathy, chronic hyperglycemia, impaired insulin signaling, and a variety of its metabolic sequelae, hyperlipidemia, and microvascular-hypoxic factors are the major mechanisms (not mutually exclusive) thought to be responsible (27; 62; 15; 48; 137).

It has been postulated that the immune system plays a role in the pathogenesis of certain diabetic neuropathies. This may be in the form of an immune-mediated attack on nerves or to their vasa nervora (as in the case of diabetics with chronic inflammatory demyelinating polyneuropathy, and proximal diabetic neuropathy respectively) (116; 88; 44). Some studies showed evidence of cytokine and chemokine production in diabetic neuropathy patients (109; 39; 86; 69). In diabetic autonomic neuropathy, this mechanism is supported by a number of observations. Lymphocytic infiltrates in sympathetic ganglia infer immune-mediated damage (40), although these findings have been questioned (118). Higher levels of TNF-α were found in type 1 diabetic patients with autonomic neuropathy compared with type 1 diabetic patients without autonomic neuropathy (109). A number of studies have reported a variety of antibodies potentially active against peripheral autonomic or somatic nerves (49).

Endoneurial microangiopathy likely plays a significant role in peripheral nerve dysfunction and pathology of diabetic polyneuropathy (51). The microvascular disease concept in diabetic sensorimotor polyneuropathy no longer addresses the complexities and unique qualities of direct neuronal involvement (155). However, the vasculitic abnormalities associated with proximal diabetic neuropathy may be the result of immune-mediated damage (116; 88; 115; 46). The response of some patients with proximal diabetic neuropathy to corticosteroids and intravenous immunoglobulin provides indirect support for an immune-mediated pathogenesis. In diabetics with chronic inflammatory demyelinating polyneuropathy, as in non-diabetics with this disorder, the mechanism is thought to be immunological, and the response to immune modulatory treatments supports this assertion (127).

The increased prevalence of focal limb neuropathies in diabetics is probably attributable to major nerve trunks being excessively prone to damage by external pressure. Animal experiments have shown that the presence of a polyneuropathy enhances the tendency to develop pressure palsies. Rats rendered diabetic with streptozotocin develop plantar neuropathies earlier and in a more severe fashion when they are kept in wire-mesh floor cages than rats that did not have diabetes (156). When a focal neuropathy occurs suddenly, the mechanism may be a nerve infarct.

Risk factors. Known risk factors for the development of polyneuropathy, in addition to hyperglycemia, are age and the duration of diabetes (105; Lasker 1993; 44). A genetic predisposition may also be important.

There is an increased body of evidence that vascular risk factors, including elevated triglyceride level, body-mass index, smoking, and hypertension, are independent risk factors for developing diabetic neuropathy (53; 44; 130).

Impaired glucose tolerance is a confirmed risk factor for developing diabetes mellitus. Patients with slightly elevated fasting glucose and HbA1c will develop frank diabetes in the next 5 years with moderate sensitivity (55%) but high specificity (98%), with specificity increasing to 99.6% at 15 years (119). This risk may be lessened with improved outcome from modest weight loss and exercise (123). It is reasonable to assume, and preliminary studies suggest, that such interventions could stabilize or improve the neuropathy. Diet and exercise counseling for impaired glucose tolerance results in cutaneous reinnervation and improved pain as confirmed by intraepidermal nerve fiber density on skin biopsy over 1 year (123).

Epidemiology

The epidemiology of diabetic neuropathies has been studied widely, but the resulting data are conflicting. The two major reasons for this are: (1) the variable criteria used for diagnosing neuropathy, and (2) the failure to recognize the different types of diabetic neuropathies (Table 1). Diagnostic criteria have included neuropathic symptoms, clinical signs of neuropathy, and electrophysiological and other quantitative laboratory abnormalities. Many studies have either concentrated on the entity of diabetic polyneuropathy or have lumped together all forms of diabetic neuropathies. It is estimated that as many as 7.7 million people in the United States have some degree of diabetic peripheral neuropathy. One third of all community-based diabetic patients in the United Kingdom have painful neuropathy symptoms, regardless of their neuropathic deficit (01). Thus, the reported prevalence of symptoms or signs of neuropathy in diabetics has varied from 10% to 100% (43).

The most accurate and informative study to date is the Rochester Diabetic Neuropathy Study (42). In this community-based study, patients were evaluated using neuropathy symptoms and physical signs and disability scores, nerve conduction studies, quantitative sensory testing, and heart rate variation studies. About 60.8% of the subjects had some form of diabetic neuropathy, although the prevalence of symptomatic neuropathy was only 14%. Of the different types of neuropathy, the most frequent by far was sensorimotor polyneuropathy (47.6%). The proportions of subjects with this were similar in the type 1 and in the type 2 groups, although the latter group had more severe manifestations of neuropathy.

In addition, patients presenting with a chronic “idiopathic” axonal polyneuropathy have nearly a twofold higher frequency of undiagnosed diabetes mellitus and impaired fasting blood glucose than age-matched controls. This suggests that an axonal neuropathy may be the presenting or the earliest manifestation in diabetes (67).

Prevention

The most important means of preventing diabetic sensorimotor polyneuropathy is optimal glucose control. This is strikingly supported by the experience with pancreas transplantation. The restoration of normoglycemia arrests the progression of chronic sensorimotor neuropathy (96). Both the Stockholm Diabetes Intervention Study and the Diabetes Control and Complications Trial showed that intensive insulin therapy reduced the development and progression of polyneuropathy in type 1 diabetics (Lasker 1993; 114). The benefits of intensive control with insulin persisted for 13 to 14 years (04). There is a similar beneficial effect on abnormal autonomic tests (10). The UK Prospective Diabetes Study reports suggest that lowering blood glucose levels in type 2 diabetics is beneficial for neuropathy (09). However, another 2-year study showed no such effect (14).

An important step in preventing complications from polyneuropathy is a rigorous and regular foot care program to screen for foot ulcers (26; 15; 95). Regular visits with a podiatrist should be encouraged.

Differential diagnosis

Confusing conditions

The mere association of neuropathic symptoms with diabetes mellitus is insufficient for the diagnosis of diabetic neuropathy; therefore, the importance of excluding other causes that might be treated differently or have a different prognosis cannot be overemphasized.

Sensorimotor polyneuropathy. The differential diagnosis is broad and includes several types of hereditary neuropathy as well as numerous causes of acquired neuropathy, including vitamin B12 deficiency, and monoclonal gammopathy of unknown significance. In practice, it is usually safe to assume that in a diabetic with a distal symmetric sensory greater than motor neuropathy, the cause is the diabetes itself. When progression of the neuropathy or proximal weakness is rapid (particularly in the upper limbs), an important consideration is chronic inflammatory demyelinating polyneuropathy.

Autonomic neuropathy. Primary and familial amyloid neuropathy can present with a combination of painful distal sensory neuropathy and autonomic failure. However, in a longstanding diabetic with these features, the overwhelming likelihood is that diabetes is the cause. Amyloid neuropathy, genetic and acquired, should be considered.

Cranial neuropathies. Although pupil-sparing third nerve palsy is a typical complication of diabetes, pupil involvement can occur in a diabetic, as well as posterior communicating aneurysm, tumors, and other mass lesions.

Truncal neuropathy. Acute or subacute onset of trunk pain and cutaneous hypersensitivity should raise the consideration of Herpes zoster, but in that condition the characteristic skin lesions should appear soon after the other symptoms. When the zone of sensory abnormalities is small, an infiltrative lesion of thoracic nerve roots should be considered and imaging studies of the spine performed.

Proximal diabetic neuropathy. Important differential diagnoses include lumbar radiculopathies from a variety of causes, including lumbar canal stenosis, carcinomatous meningitis, and malignant invasion or other mass lesion (hematoma or abscess) of the lumbar plexus.

Associated or underlying disorders

Chronic inflammatory demyelinating polyneuropathy (CIDP). There is controversy as to whether chronic inflammatory demyelinating polyneuropathy occurs more frequently in diabetics than in nondiabetics (79; 127; 71; 59; 35; 120; 81). Although one study suggested a reduced risk of CIDP with diabetes (81), another study indicated a two-fold increased relative risk of diabetes compared with the general population (113). The diagnosis of chronic inflammatory demyelinating polyneuropathy in diabetic patients remains difficult, and it should be clinically based on the typical CIDP phenotype first and foremost and then with electrodiagnostic support (136).

Focal limb mononeuropathies. In most diabetics with focal neuropathies involving the major nerves in the limbs, the underlying causes are similar to those of nondiabetics.

Diagnostic workup

Diagnostic evaluation for suspected diabetes and impaired glucose tolerance. In the case of a patient not known to be diabetic but presenting with a neuropathy, fasting blood sugar and glycosylated hemoglobin levels are often sufficient to make the diagnosis of diabetic neuropathy. If these are negative, an oral glucose tolerance test should be done to evaluate for impaired glucose tolerance (107). Evaluating for features of the metabolic syndrome should also be considered. Intraepidermal nerve fiber density assessed by skin punch biopsy (106) or by skin blister (99) is an emerging technique that is gaining popularity in the diagnosis of small fiber neuropathy and should be included as an endpoint in neuropathy trials. A high index of suspicion is required to diagnose diabetic autonomic neuropathy, and a validated self-reported clinical questionnaire (the Survey of Autonomic Symptoms) improves diagnostic sensitivity (154).

Diagnostic evaluation of diabetic neuropathies. Standard motor and sensory nerve conduction studies and needle electromyographic examination of muscles are the basic techniques used for evaluating the various types of diabetic neuropathies (103). Because nerve conduction studies and needle electromyography evaluate the large diameter myelinated fibers only, they can be misleadingly normal in patients with polyneuropathy in whom the involvement is mainly of the small diameter nerve fibers (small fiber neuropathy). Techniques for evaluating that population of peripheral nerve fibers are discussed below.

Sensorimotor, small fiber, and autonomic polyneuropathy. Several diagnostic modalities are available to study nerves in diabetic neuropathy (Table 2).

Table 2. Diagnostic Modalities Available to Study Nerves in Diabetic Neuropathy

CASE IV = computer-assisted sensory examination
QSART = quantitative sudomotor axon reflex test
NCS = nerve conduction studies
IENFD = intraepidermal nerve fiber density

Electrodiagnosis. The electrodiagnostic features of diabetic sensorimotor peripheral polyneuropathy are characteristic of a primarily axon loss sensorimotor polyneuropathy with some evidence of demyelination, which can be quite prominent but insufficient in isolation to make a diagnosis of CIDP. The lower limbs are affected first, so the priority should be for nerve conduction studies to be done there. Because it is usually a symmetrical process, one limb can be studied. The tibial and the peroneal motor nerve conduction studies are performed, recalling that the peroneal motor response recording extensor digitorum brevis is prone to focal muscle damage. The sural nerve and the superficial peroneal sensory nerve actions potentials are essential in evaluating large fiber sensory fibers. Tibial H-reflexes and F waves are useful adjunctive studies. To evaluate if the neuropathy is severe enough to involve the upper limbs, median, ulnar, or radial nerves can be studied. The radial sensory nerve conduction study is preferred because of the propensity for the other nerves to suffer focal nerve entrapment at the elbow or wrist.

Early changes are typically restricted to the lower extremities and consist of one or more of these NCS abnormalities: absent H-reflexes, low amplitude or absent sural and superficial peroneal sensory responses, low amplitude tibial and peroneal motor responses, mild slowing of sensory and motor distal latencies and conduction velocities, and active denervation or reinnervation in distal leg muscles. With more advanced disease, the upper extremities become involved. This often manifests by reduction of the median, ulnar, and radial sensory amplitudes and low or borderline median and ulnar motor amplitudes with mild sensory and motor conduction slowing. In severe diabetic polyneuropathy, there is often complete absence of all routine sensory and motor conduction studies in the legs and absent sensory responses in the hands with very low amplitude median and ulnar motor responses in the upper limbs. The needle electromyography often shows long-duration, high-amplitude, and rapidly recruited motor unit action potentials with or without fibrillation potentials. This is usually symmetrical and worse in the leg muscles distally with a clear distal to proximal gradient.

Neuromuscular ultrasound. Imaging of peripheral nerve using high resolution ultrasonography is increasingly used in clinical practice. Peripheral nerves are enlarged diffusely in diabetics, including those with diabetic polyneuropathy, when compared to controls (20). However, as with other axonal polyneuropathies, the enlargement or nerve morphology changes in patients with diabetic polyneuropathy are often mild with considerable overlap in nerve size between patients and control subjects (128).

The utility of ultrasound in diabetic patients is twofold. First, it is useful in the diagnosis of nerve entrapment syndromes by confirming focal nerve focal cross-sectional area (CSA) enlargement at the entrapment sites, which has been the most commonly reported indicator of nerve pathology. Second, neuromuscular ultrasound may be able to distinguish the axonal nature of diabetic polyneuropathy from the acquired and inherited demyelinating polyneuropathies such as CIDP or Charcot-Marie-Tooth disease (58; 128).

Quantitative sensory testing. Techniques for detecting small nerve fiber involvement include quantitative sensory testing (126; 131). In quantitative sensory testing, graded degrees of warm, cool, or vibration sensations can be applied to the patient's extremities; the thresholds for sensory perception are established and compared with normal values. Temperature sensation is mediated via small diameter nerve fibers, so quantitative sensory testing is a useful complimentary test to nerve conduction studies. Vibration is mediated through large diameter sensory fibers, so testing this modality overlaps with sensory nerve conduction studies. The use of quantitative sensory testing in diabetic neuropathy is specifically reviewed by Chong and Cros (33). Abnormal thresholds for cool and warm sensations may be detected when the nerve conduction studies are normal, confirming the presence of a small fiber or a predominantly small fiber neuropathy. Questions have been raised as to the reproducibility (test-retest reliability) of thermal threshold tests that might limit their use in following the course of neuropathy as in a drug trial study (135). This does not, however, detract from its use as a diagnostic tool in the patient with neuropathic symptoms and normal nerve conduction studies.

Autonomic testing. Specialized sudomotor evaluation techniques include heart rate variability, quantitative sudomotor axon reflex testing, thermoregulatory sweat testing, and sweat droplet testing (31). Noninvasive cardiovascular procedures, such as heart rate variability with deep breathing or Valsalva maneuver, are relatively sensitive methods for detecting early cardiovagal denervation. A resting tachycardia and a fixed heart rate on deep breathing and going from lying to standing can be demonstrated by EKG recordings. Noninvasive recording of beat-beat blood pressure can be done in specialized laboratories (89; 31). Sudomotor function tests, including thermoregulatory sweat testing and quantitative sudomotor axon reflex test (QSART), may reveal abnormalities. Hemodynamic tilt tests often are valuable when orthostatic hypotension is present.

Several autonomic tests have been recommended (Level B) by the American Academy of Neurology for evaluation of patients with distal symmetric polyneuropathy, including patients with diabetic neuropathy (50). These include:

Gastrointestinal autonomic dysfunction and atony is assessed with various radiographic techniques, but the easiest is simply to demonstrate the abnormally slow passage of barium through the gut (92). To accurately define the nature of neurogenic urinary bladder and sphincter dysfunction, a battery of specialized tests is required. Urinary bladder dysfunction can be investigated by quantitating the post-voiding residual urine, which typically is increased in diabetic cystopathy; cystoscopy and urodynamic studies are required to document and quantitate the degree of bladder dysfunction. Male sexual dysfunction, including erectile impotence, can be differentiated from psychogenic impotence with various tests, including penial tumescence studies during REM sleep.

The nerve fibers that mediate sweating undergo distal damage in polyneuropathies. One electrophysiological technique for evaluating these nerve fibers is to test for sympathetic skin responses. This can be done with most standard EMG machines (97). The utility of this technique for the diagnosis of diabetic neuropathy and diabetic autonomic neuropathy has been questioned (23).

Skin biopsy. This is a technique of distal leg skin punch biopsy with quantification of the epidermal nerve fiber density (66; 63; 75). This is particularly useful in confirming small fiber neuropathy, diabetic or otherwise (82). Intraepidermal nerve fiber density assessing skin blister is likely equivalent to skin punch biopsy (99). Small fiber neuropathy is closely linked to the metabolic syndrome, which consists of diabetes/prediabetes, hypertension, dyslipidemia, and central obesity (150).

Cutaneous nerve biopsy. This is rarely indicated in patients with sensorimotor or other diabetic neuropathies. It may be of value in the unusual circumstance when the differential diagnoses of amyloidosis or vasculitis are being considered.

Corneal confocal microscopy. Reduced corneal sensation has long been known to occur in patients with diabetic polyneuropathy. The cornea, one of the most densely innervated tissues of the human body, showed reduced sensation symmetrically early in diabetes and this correlates with disease duration (37). Corneal nerve fiber and nerve branch density significantly correlates with intraepidermal nerve fiber density and QST in patients with small fiber neuropathy (72).

Focal limb neuropathies. When encountered in diabetics, carpal tunnel syndrome, ulnar neuropathies at the elbow, and other focal neuropathies are often superimposed on a diabetic sensorimotor polyneuropathy. Nerve conduction studies and needle electromyographic examination can be used to evaluate the presence and severity of such focal neuropathies. Useful strategies to distinguish such focal neuropathies from diabetic polyneuropathy include using internal comparison studies in the hand such as the median-ulnar palmar mixed studies, median-ulnar sensory studies recording ring finger, median-radial sensory studies recording thumb, and median-ulnar comparative motor studies recording second lumbrical-interossei (133). Among these studies, the median-ulnar comparative motor studies recording 2nd lumbrical-interossei is the most useful in patients with suspected carpal tunnel syndrome and superimposed on moderate or severe diabetic polyneuropathy where the sensory studies are either very low or absent (133). More stringent criteria (greater than 0.8 msec vs. greater than 0.5 msec latency difference) are often needed for accurate confirmation of carpal tunnel syndrome in patients with underlying peripheral polyneuropathy.

Diabetic radiculoplexopathy. In patients with this type of neuropathy, nerve conduction studies usually show the presence of distal sensorimotor polyneuropathy. No nerve conduction tests are specifically useful for the diagnosis of diabetic radiculoplexopathy, but reduction of femoral compound action potential amplitude on side-to-side comparison is a helpful clue when present. More helpful is needle electromyographic examination, which usually shows asymmetric neurogenic abnormalities in proximal muscles most often affecting the quadriceps, iliacus, and thigh adductor muscles. These needle electromyographic examination abnormalities may extend as far proximally as the paraspinal muscles. It is important to perform imaging studies of the lumbar spine, the retroperitoneal area, and sometimes the sacrum and pelvis; caution must be used in the diagnosis of compressive radiculopathy in this setting. Cerebrospinal fluid analysis is indicated if there is a suspicion of carcinomatous meningitis producing polyradiculopathy or chronic inflammatory demyelinating polyneuropathy.

Truncal (thoracic) neuropathy. Needle electromyographic examination studies can show neurogenic abnormalities in paraspinal and abdominal muscles at the segmental level of the sensory abnormalities determined clinically. Thermoregulatory sweat testing is useful in confirming the sensory deficit by showing loss of sudomotor function in the involved thoracic spinal nerves. Imaging of the thoracic spine should be considered when the area of sensory symptoms is small, or if there is a lot of pain in the spine itself.

Coexisting chronic inflammatory demyelinating polyneuropathy. The diagnosis of CIDP should be clinically based on the typical CIDP phenotype first, after which one can consider whether there is electrodiagnostic supportive evidence (136). Nerve conduction studies and even nerve biopsy unfortunately sometimes fail to distinguish diabetic symmetrical polyneuropathy from chronic inflammatory demyelinating polyneuropathy (127). Cerebrospinal fluid analysis for elevated protein levels is not helpful in this situation either, as an elevated protein may be found in diabetics without neuropathy (127). Neuromuscular ultrasounds are helpful by showing focal and regional enlargements in proximal nerves of the upper extremities and brachial plexus (128). In addition, the positive response to immunotherapy may provide additional evidence for chronic inflammatory demyelinating polyneuropathy.

Management

Supportive measures are discussed in Peripheral neuropathies: supportive measures and rehabilitation.

Sensorimotor polyneuropathy. The Stockholm Diabetes Intervention Study, the Diabetes Control and Complications Trial, and other studies have shown that tight glucose control can slow the progression of polyneuropathy in type 1 diabetics (08; 114; 70). This tight control extends its protective effect for at least 10 years (04). The UK Prospective Diabetes Study addressed the same question for type 2 diabetics but using different and perhaps less rigorous measures of polyneuropathy (09). The results suggest some benefit after long term intensive glucose control. Pancreas transplantation generally causes the polyneuropathy and diabetic autonomic neuropathy to stop worsening; the former may improve slightly (96). Studies suggest that supervised exercise results in improvements in neuropathic symptoms, neuropathic pain, and cutaneous nerve fiber branching in patients with diabetic peripheral neuropathy (76).

Enormous effort has been put into the search for specific treatments, but the results have been disappointing (17). Such treatments include myoinositol, essential fatty acids, vitamins, protein kinase C inhibitors, vasodilators, antiprostaglandins, ACE inhibitors, lipid lowering agents, advanced glycation end product inhibitors, acetyl-L-carnitine gangliosides, neurotrophic factors, and a variety of aldolase reductase inhibitors (21). The most promising of these in animal models of diabetic sensorimotor neuropathy are the aldolase reductase inhibitors, but they have been generally ineffective in humans, and some have important side effects (03; 21; 52). Alpha-lipoic acid, an antioxidant, has shown promise in some clinical trials and continues to be evaluated in other trials (153; 07). Seal oil omega-3 polyunsaturated fatty acids supplementation increased corneal nerve fiber length in patients with distal sensorimotor polyneuropathy and type 1 diabetes mellitus (84).

Two issues deserve emphasis regarding attempts at treating diabetic sensorimotor neuropathy. The damage may be too advanced by the time the treatment attempts are started, so early intervention is important. It would also seem important to consider multifactorial interventions, including behavior modification (more exercise, better diet, and smoking cessation) and pharmacological therapy, targeting not only hyperglycemia but hypertension, dyslipidemia, and microalbuminuria. Such an approach has been shown to slow progression of retinopathy, nephropathy, and autonomic neuropathy. Disappointingly, there was no impact on sensorimotor neuropathy (54).

Hyperglycemic neuropathy, treatment-related neuropathy, and acute painful sensory neuropathy eventually resolve following the establishment of euglycemia.

Diabetic autonomic neuropathy. There is evidence from both the Diabetes Control and Complication Trial and the UK Prospective Diabetes Study that long-term intensive glucose control improves measures of autonomic function in both types of diabetics (09; 10). It appears that intensive glucose control does not reduce the risk of developing erectile dysfunction (09).

The main requirement in patients with autonomic neuropathy is symptomatic treatment that focuses on the involved organs or systems. Orthostatic hypotension is managed by removing aggravating medications, when possible, instituting nonpharmacological maneuvers such as the head-up tilt of the bed and elastic stockings (98). Effective medical therapy includes fludrocortisone, midodrine, pyridostigmine, and off-label use of droxidopa based on case reports (146; 98; 74). Gastroparesis is treated with dietary manipulation and prokinetic and antiemetic medications; diarrhea and constipation are treated with the medications used in nondiabetics with these symptoms (92). About 50% of diabetic men with erectile dysfunction will respond to sildenafil, tadalafil, or vardenafil. If these fail, intraurethral or intrapenile vasoactive agents, a vacuum constriction device, and penile prostheses are other options (147). Bladder dysfunction requires urologic investigations to define the nature of the dysfunction, which then acts as a guide to management regarding the choice of medications, catheterization, or surgery (147).

Painful third nerve palsy requires analgesics and patching the eye to eliminate diplopia. The management of truncal neuropathy is mainly that of controlling neuropathic pain. Carpal tunnel syndrome should be treated in the same way as in nondiabetics (34). The management of other focal neuropathies in diabetics, such as ulnar neuropathies at the elbow, is less clear.

An open-label uncontrolled study on the treatment of proximal nondiabetic neuropathy with intravenous methylprednisolone had initially produced promising results (45). Multicenter controlled studies are needed to evaluate the efficacy of this medication and that of intravenous immunoglobulin in diabetic radiculoplexopathy (46). A Cochrane review (29) reviewed the only controlled trial in diabetic lumbosacral radiculoplexus neuropathy, which was unpublished (47). This review concluded that there was no evidence to support a positive or negative effect of any immunotherapy in the treatment of diabetic amyotrophy.

Options for treating the rare case of confirmed chronic inflammatory demyelinating polyneuropathy in a diabetic based on clinical phenotype first followed by electrodiagnostic support includes various prednisone, intravenous methylprednisolone, intravenous immunoglobulin, or plasma exchange as first-line therapy. Other options include azathioprine, mycophenolate mofetil, cyclosporin (79; 127; 71; 136). One small study compared intravenous immunoglobulin with plasmapheresis and found no difference; there was no additional effect when prednisone was added to either (71). All patients in these three studies improved, usually impressively so. The response onset was anywhere from 2 weeks to 4 months and reached its maximum anywhere from 1 to 12 months. There were no recurrences. Another study reported only modest beneficial effects from intravenous immunoglobulin treatment (35).

Pain management. In many of the diabetic neuropathies, a prominent feature is neuropathic pain. The details of the management of neuropathic pain are discussed in other sections; however, a summary of the medications categorized into various levels based on the presence of randomized studies and the level of evidence available for use in the management of neuropathic pain is presented in this section (102). Table 3 and Table 4 are presented for guidelines (143; 18; 15; 138; 102). An evidence-based guideline from 2011 suggested that pregabalin is effective and should be offered for the relief of pain in diabetic neuropathy (Level A), whereas venlafaxine, duloxetine, amitriptyline, gabapentin, valproate, opioids (morphine sulfate, tramadol, and oxycodone controlled-release), and capsaicin are probably effective and should be considered for treatment of painful diabetic neuropathy (Level B) (22). An update to this practice guideline was published in Neurology in December 2021. According to this guideline, the clinicals should be assessed for mood and sleep disorders (Level B). Patients with diabetic neuropathic pain should be offered tricyclic antidepressants, serotonin norepinephrine reuptake inhibitors, or sodium channel blockers to reduce pain (Level B). Consider factors other than efficacy (Level B), its effects on function, and quality of life. Consider medication from different classes when patients do not achieve meaningful improvement or if they experience significant adverse effects with the initial therapeutic class (Level B). The use of opioids for the treatment of painful diabetic neuropathy is not recommended (Level B) (110).

Lidocaine patch or cream may be used as an adjunctive treatment; it blocks sodium channels and reduces the hyperexcitability of nociceptors (145). Isosorbide dinitrate spray, a nitric oxide-dependent vasodilator, showed benefit in reducing burning sensation and neuropathic pain (148). There is increasing evidence showing some benefit of alpha lipoic acid in patients with diabetic neuropathy (151; 94; 100). Aldose reductase inhibitor, epalrestat, has been found to reduce neuropathic symptoms (68).

Topical therapy for painful diabetic neuropathy has mainly centered on the use of lower concentration capsaicin cream. A meta-analysis provides some evidence for its efficacy, although true blinding of the studies on which this analysis was made is questionable due to the hyperalgesia following the application of the cream (149). Some patients find this initial hyperalgesia intolerable, and it requires care in its application so that the cream does not get in the patients' mouths or eyes. It is only recommended for use for up to 8 weeks at a time.

In 2020, the 8% capsaicin patch was approved by the FDA for neuropathic pain associated with diabetic peripheral neuropathy (https://www.accessdata.fda.gov/drugsatfda_docs/label/2020/022395s019lbl.pdf). Results of the 12-week double-blind placebo-controlled trial showed a 30% change in average pain from baseline to week 12 in patients who were treated with the 8% capsaicin patch compared to 22% with placebo. Application site erythema, pain, and pruritus were seen as side effects without any major side effect (122). One controlled study reported that the local application of isosorbide dinitrate spray is helpful in reducing diabetic neuropathic pain (148). Studies support the use of topical capsaicin, glyceryl trinitrate spray, Citrullus colocynthis, nontraditional therapies (Ginko biloba), and other therapies, such as exercise, cognitive behavioral therapy, and mindfulness (110).

In patients with refractory painful diabetic neuropathy, spinal cord stimulation has been shown to provide pain relief and improved quality of life (104; 112). In July 2021, the FDA approved the spinal cord stimulator for pain associated with diabetic neuropathy in the lower limb.

Table 3. Medications for Painful Neuropathy

Table 4. Guidelines for Medication Use in Treating Peripheral Neuropathic Pain

With the increased prevalence of chronic nerve entrapments in patients with diabetes mellitus, several authors have proposed that surgical decompression of distal lower extremity peripheral nerves are possibly beneficial in patients with painful diabetic polyneuropathy, as well as in patients with painful cryptogenic polyneuropathy (134). Common sites of proposed decompressions include the tibial nerve at the ankle (including calcaneal, medial, and lateral plantar nerves in the foot), deep fibular (peroneal) nerve at the ankle, common fibular (peroneal) nerve at the knee, lateral femoral cutaneous nerve at the inguinal ligament, and sural nerve in the posterior calf region. However, the role of decompressive surgery for diabetic symmetric distal neuropathy is unproven because there are no well-designed randomized controlled trials (30). A single center randomized study of 42 diabetic patients showed improved pain in the surgical treated foot compared to the contralateral foot and to baseline based on visual analogue scale (91). The Study on the Role of Decompression of Lower Extremity Nerves for the Treatment of Patients With Symptomatic Diabetic Neuropathy With Chronic Nerve Compression was completed in 2013, but no results have yet been posted (https://clinicaltrials.gov/ct2/show/NCT00703209). This intervention remains to be controversial in diabetic neuropathy, and we caution patients against it until evidence from multicenter controlled trials definitively demonstrates any benefit.