Diabetes Mellitus: A Fundamental and Clinical Text
3rd Edition

Numerous population studies of diabetes mellitus and its long-term complications support the idea that there is an association between diabetes and oxidative stress (53,54,55). Less certain, however, is whether oxidative stress contributes to the development of long-term complications or merely reflects associated processes that are affected by diabetes. Diabetic neuropathy is among the complications recognized to be associated with increased oxidative stress (56,57). An increase in oxidative stress may occur because of either an increase in free radical production or a reduction in antioxidant defenses (58). There are many suggestions regarding the origins of oxidative stress in diabetes, including free radical accumulation related to glycation of proteins (59), consumption of NADPH through the polyol pathway (60), glucose autoxidation (61), hyperglycemia-induced pseudohypoxia (62), or activation of PKC (63). Diabetic monocytes also have an increased capacity to produce superoxide anion (64). In addition, superoxide dismutase, which has the important role of neutralizing superoxide radicals, is reduced in diabetic peripheral nerve tissue, thus compounding any enhancement of free radical formation (65,66). When not scavenged properly, superoxide anion may interact with NO produced by vascular endothelium or nitrergic nerves to form peroxynitrite (ONOO–) (67), which would result in the reduction of endothelium-dependent vasorelaxation and nitrergic neurotransmission. Peroxynitrite can readily nitrosylate proteins, potentially altering their function, or break down to give highly reactive hydroxyl radicals that are cytotoxic (68). Increased concentrations of nitrotyrosine have been noted in vessels from diabetic patients and endothelial cells cultured under high glucose conditions (69). Furthermore, peroxynitrite can avidly oxidize tetrahydrobiopterin, an endothelial nitric oxide synthase (eNOS) cofactor, to dihydrobiopterin. Under conditions of BH₄ deficiency, eNOS is in an uncoupled state, resulting in the production of superoxide, rather than NO. Indeed, there is direct evidence for a dysfunctional, uncoupled NOS in experimental (70) and clinical studies (71), showing that the administration of the eNOS cofactor BH₄ improves endothelial dysfunction in the setting of diabetes mellitus. The peroxynitrite formation can be enhanced in diabetes not only by increased superoxide formation, but also by enhanced NO formation, due to upregulation of NOS in some tissues, particularly early in the disease (72). Endothelium-dependent relaxation of aorta from diabetic rats could be protected by NO scavenger treatment (73). Also, the early degenerative changes of corpus cavernosum nitrergic innervation were partially attenuated by NO synthase inhibitor treatment (74). However, the widely accepted
strategy in improving the oxidative balance in diabetes does not target the reduction of NO production, but rather inhibits the formation of reactive oxygen species, including superoxide anion and efficiently scavenging peroxynitrite. In general, both lipophilic and hydrophylic scavengers are effective against endothelial and nerve dysfunction in experimental diabetes. Another approach is to prevent reactive oxygen species formation by autoxidation and the Fenton reaction, which are both catalyzed by free transition metals, particularly iron and copper. The normally tight regulation of free transition metals is compromised in diabetes (75). Clinical neuropathy trials in diabetic patients are under way using lipoic acid, which is both a scavenger and a metal chelator. Early reports show some improvement for autonomic neuropathy as well as modest improvements in nerve conduction velocity (76).

Recent evidence in experimental animals has indicated that hyperglycemia stimulates the production of NO, which reacts with superoxide anion to form peroxynitrite, which is damaging to the endothelium of blood vessels (77) and the perineurium of nerves (78). This at first glance appears contradictory to the abundant data indicating that NO is a beneficial endothelium-derived vasodilating factor that is deficient in diabetes (77). However, NO production in the endothelium mediated by eNOS may respond differently to NO production in the skin ostensibly mediated by inducible or neuronal nitric oxide synthase (iNOS or nNOS). NO is produced by keratinocytes, macrophages, and smooth muscle cells and nitrergic neurons in the skin, and it has been shown (77) and it has been reported that nitrotyrosine (NTY) staining of the endothelium in diabetic rats is associated with failure of endothelial-dependent vasodilation (77). The findings of normal or overproduction of NO in the skin of patients with diabetes (79) are in keeping with the notion that this is a compensatory mechanism for the impaired vasodilation or that it reflects a pathologic overproduction of an alternative source of NO. Stimulation of iNOS and NO overproduction also enhances lipid peroxidation, and the formation of lipid peroxides in nerve membranes has adverse effects on nerve function, membrane fluidity, and electrical activity (80). Most recently Hoeldtke et al. measured circulating levels of plasma nitrite and nitrate (collectively NOx) since peroxynitrite is relatively unstable in 37 type 1 diabetics and 41 healthy controls and showed elevated levels of NOx and NTY in the diabetic patients. These patients were monitored for 3 years, and positive correlations were found for NCV in the median, ulnar, and peroneal nerves in patients with high levels of NTY (81). NO overproduction in diabetes has now been reported in several studies in animals (26) and clinical studies in humans (82,83). It is therefore not inconceivable that the NO overproduction in skin reflects this nitrosative stress, and the correlation we have found in nerve function and the excretion of 8-iso-PGF2α is supportive of this concept. The origin, however, of the NO remains to be determined.

Metabolic factors cannot account for all forms of neuropathy, or for the heterogeneity of the clinical syndromes. In a subpopulation of neuropathic patients, immune mechanisms may be responsible for the clinical syndrome, especially in patients with the proximal variety of neuropathy and those with a more marked motor component to their neuropathy. Our data support the hypothesis that circulating antineuronal antibodies are present in diabetic serum, at least in some patients (84,85). The circulating autoantibodies directed against motor and sensory nerve structures have been detected by indirect immunofluorescence, and antibody and complement deposits in various components of sural nerves have been shown (86,87,88).

A frequent (12%) association of antiganglioside (GM1) antibodies with distal symmetric peripheral neuropathy characterized by a slight emphasis on a motor deficit with electrophysiologic signs of demyelination was found in our population studies, but we are a tertiary referral center, and this may not be true for neuropathy in general. We have also found antiphospholipid antibodies (PLAs) in 88% of our diabetic population with neuropathy, compared with 32% in diabetic patients without apparent neurologic complications, and in only 2% in the general population (89). PLAs are present in a number of autoimmune, neurologic, and hematologic disorders (90). There is evidence that PLAs may indeed be injurious to neural tissue and that damage may be selective for specific parts of the nervous system (89,90). Because PLAs are associated with a tendency to develop vascular thrombosis, their presence may provide a link between the immune and vascular theories of causation of neuropathy. Because our understanding of autoimmune neuropathies in general is constantly being fueled by new evidence, this area of research in DM promises to be exciting and fruitful.

Microvascular insufficiency has been proposed by a number of investigators as a possible cause of diabetic neuropathy (91,92,93). The interest in microvascular derangement in diabetic neuropathic patients has arisen from studies suggesting that absolute or relative ischemia may exist in the nerves of diabetic subjects because of altered function of the endoneurial or epineurial blood vessels. Histopathologic studies show the presence of different degrees of endoneurial and epineurial microvasculopathy, mainly thickening of the blood vessel wall or occlusion (94,95). A number of functional disturbances have also been demonstrated in the microvasculature of the nerves of diabetic subjects. Studies have demonstrated decreased neural blood flow (96), increased vascular resistance (97), decreased PO₂ (52,96), and altered vascular permeability characteristics such as a loss of the anionic charge barrier and decreased charge selectivity. It has also been shown that abnormalities of cutaneous blood flow correlate with neuropathy (98,99), suggesting that there is a clinical counterpart to the microvascular insufficiency that may prove to be a simple noninvasive test of small nerve fiber dysfunction.

Persistent hyperglycemia increases polyol pathway activity with accumulation of sorbitol, fructose, and advanced glycation end products (AGEs) in nerves, damaging them by an as yet unknown mechanism (44,51) (Fig. 91.2). This is linked to altered activity of PKC, in all likelihood the β2 moiety, implicated in the pathogenesis of neuropathy by as yet undefined mechanisms. Alternatively, a hyperglycemia-induced increase in diacylglycerol levels leads to the activation of PKC, which then modulates the activity of Na⁺/K⁺-ATPase (100) in both neurons and Schwann cells (SCs). Neural and glial elements of the peripheral nervous system maintain their homeostasis by a bidirectional interaction, either as direct contact or through the
local release of autocrine and paracrine soluble mediators such as the various cytokines. Changes in the activity of PKC or Na⁺/K⁺-ATPase alter the expression of a variety of genes, including the cytokines. It has been shown that impaired activity of Na⁺/K⁺-ATPase or increased activity of PKC leads to the upregulation of the inflammatory cytokines interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α) gene expression in mononuclear cells (101). PKC-mediated intercellular signaling is one of the major pathways of information transfer through cytokine receptors in glial cells. Cytokines can in turn modulate the activity of Na⁺/K⁺-ATPase (102,103), suggesting that the changes in the sodium pump may be a consequence of cytokine activation rather than a cause. The situation, however, is not all that clear because SC proliferation, which is necessary for nerve regeneration after injury, is also mediated through PKC activity (104,105), and it remains to be determined whether PKC exerts harmful or beneficial effects in the pathogenesis of neuropathy and attempts at repair. Deficiency of GLA as well as N-acetyl-L-carnitine have also been implicated (106). It is also apparent that oxidative stress with a reduction in available endogenous antioxidants plays a role in aging and its effects on the peripheral nervous system. Nitric oxide (NO) has been suggested as the potential bridge between the metabolic and the vascular hypotheses (107). One of the actions of cytokines is the induction of iNOS and the production of NO. SCs have been shown to upregulate iNOS and produce NO in response to cytokines (108). In endothelial cells, a decrease in Na⁺/K⁺-ATPase synergizes the effects of proinflammatory and immune cytokines on iNOS induction as well as the induction of various cell adhesion molecules (CAMs) (109). The relative inefficacy of aldose reductase inhibitors (ARIs) (110) and the failure of intense glycemic control to abolish the development of neuropathy have suggested that alternative mechanisms for the pathogenesis of neuropathy must be sought. We have shown that circulating levels of the cytokine IL-6 are higher in diabetic patients with neuropathy than in non-neuropathic diabetic patients, and Jude et al. (111) have shown that there is a predictive value of elevated levels of the CAM p-selectin for the development or progression of diabetic neuropathy or both, independent of glycemic control.

Apart from the metabolic, immunologic, and vascular factors involved in the pathogenesis of neuropathy, there are data to support a role for growth factor deficiency. Many of the neuronal changes characteristic of diabetic neuropathy are similar to those observed after either removal of target-derived growth factors by axotomy or depletion of endogenous growth factors by experimental induction of growth factor autoimmunity. Because neuronal growth factors can promote the survival, maintenance, and regeneration of neurons subject to the noxious effects of DM, as well as affect their function, the success of diabetic patients in maintaining normal nerve morphology and function may ultimately depend on the expression and efficacy of these factors (106).

It has been known for some time that sympathetic and dorsal root ganglion (DRG) neurons are developmentally dependent on nerve growth factor (NGF), but, more recently it has been shown that adult DRG and sympathetic neurons, both populations of neurons affected in diabetic neuropathy, depend on NGF either for their maintenance or their survival (111). NGF belongs to a gene family encoding structurally and functionally related proteins called neurotropins. NGF has been implicated in diverse and widespread activities, including vasodilatation, gut motility, and nociception (111). There are data to suggest that a decline in NGF synthesis in DM has a role in the pathogenesis of neuropathy, namely, in functional deficits of small fibers. These fibers have a role in pain and thermal sensation. The effect of NGF depletion may be mediated through the downregulation of neurofilament gene expression, or messenger
ribonucleic acids that encode the precursor molecules of substance P, both shown to be NGF dependent (106). Another member of the neurotropin family, neurotropin 3, may be important for the survival and function of the large nerve fibers subserving position, vibration, and possibly motor functions (106). The insulin-like growth factors (IGFs), IGF-1 and IGF-2, have also been implicated in the growth and differentiation of neurons, and IGF receptors are present in nerve tissues (i.e., neurons, SCs, ganglia) involved in DM-associated nerve disorders. IGFs and their binding proteins are regulated by insulin and the glycemic state (106,112,113). One of the consequences of insulin insufficiency is reduction in circulating IGF-1 concentration. It seems reasonable to hypothesize that abnormal IGF-1 and IGF-2 metabolism plays a causative role in some aspect of diabetic neuropathy. However, little is known about the other effects of DM on local expression, synthesis, and transport of these growth factors in the nerve tissue.

Schwann cell proliferation is essential for the successful regeneration of injured peripheral nerve (104,105). Morphologic changes in SCs such as onion bulb formation have been well described in both animal models and human DM (114), and biochemical defects in SCs related to the extracellular matrix proteins (ECMs) and insufficient production of growth factors or overproduction of inflammatory cytokines have been implicated (115,116,117).

Regeneration and remyelination of the peripheral nerve is mediated by SC-axonal contact through extracellularly derived ECMs, which form the basal lamina, and corresponding SC membrane ligands, CAMs. The necessary signals for these processes may be provided by direct contact between SCs and axons, or by soluble mediators such as the cytokines (118). Cytokines modulate important glial cell functions, such as the induction of growth factors, other cytokines, ECMs, and CAMs, and protection from apoptosis (119). We have shown that IL-6, for example, is able to overcome in part the neurotoxicity of serum of patients with type 1 DM on neuroblastoma cells in vitro. Thus, the pleiotropic cytokines can function in both neurodegeneration as well as regeneration and neuroprotection (120,121). When nerves are damaged by trauma, toxins, or infections, cytokine production by SCs, macrophages, mast cells, and neurons is increased. These include the proinflammatory cytokines (e.g., IL-1, IL-6, and TNF-α), the immunomodulatory cytokines such as TGF-β, and the immune cytokines, such as IL-2 and interferon-γ. By their vast array of actions through autocrine, paracrine, and direct cellular effects, these cytokines can exert major effects on SCs with regard to ECM production, CAM expression (122,123), production of trophic and growth factors (124,125), and production of the cytoskeleton (126,127).

In the peripheral nerve, three anatomic compartments with different cellular and ECM composition can be distinguished: epineurium, perineurium, and endoneurium. The epineurium surrounds the perineurial ensheathment of the fascicles and contains collagen type 1, fibroblasts, fat cells, blood vessels, and lymphatics. The perineurium consists of flattened perineurial cells covered by basement membrane and interspersed collagen fibrils and tenascin c. These cells are linked together with tight junctions that constitute the blood–nerve barrier. The endoneurium contains collagen types 1 and 3 and encases the blood vessels and the axon–SC units, macrophages, and fibroblasts. The basement membrane is composed of ECM proteins secreted by the SCs. Depending on the nature of the neuritis, T cells and immunoglobulins are found with these different components, and there may be characteristic aspects of the different neuropathies.

Experimental allergic neuritis (EAN) is an acute demyelinating inflammatory neuropathy that is induced in susceptible species by active immunization with peripheral myelin proteins or by adoptive transfer of neuritogenic T cells. Integrin expression on SCs is associated with defined histopathologic alterations in EAN that resemble the findings in human neuropathies. In vitro, both cytokines and the assembly of the basal lamina influence the expression pattern of integrins on SCs. The observed neoexpression of α4β5 in EAN is important because in early development it is strongly present on SCs, is downregulated in the adult, and is again neoexpressed in noninflammatory conditions such as nerve regeneration after nerve transection in the chick. However, other integrins may play a vital role in the development of neuritis.

Laminin is a large, heteromeric, cruciform glycoprotein composed of one large A chain and two smaller B chains, β1 and β2 (128,129). Laminin has been shown to promote neurite extension by cultured neuroblastoma cells (130,131,132,133,134). Our studies show that lack of normal expression of the laminin β2 gene may contribute to the pathogenesis of diabetic neuropathy. First, we showed that all normal adult rat DRG neurons express the laminin β2 chain gene (135). In other experiments, we described the upregulation of the laminin β2 gene in regenerating neurons (136,137). This finding indicates its importance in the process of neuronal regeneration. However, laminin β2 expression under basal conditions and during neuronal regeneration was decreased in diabetic animals (138). If these changes are present in diabetic patients with neuropathy, it might lay the foundation for exploring the therapeutic potential of laminin.

In summary, diabetic neuropathy is a heterogeneous disease with widely varying pathologic effects, suggesting differences in the pathogenetic mechanisms of the different clinical syndromes. Recognition of the clinical homologues of these pathologic processes is the first step in achieving the appropriate form of intervention.

Foot ulcer constitutes a major source of morbidity among patients with DM. Loss of protective sensation and repetitive trauma (e.g., walking) are the major causes. Ulceration occurs most frequently over the metatarsal heads, but also appears at other areas of increased pressure. Loss of tone in the small muscles of the feet leads to an imbalance between the flexors and extensors, ultimately resulting in the classic hammer-claw toe. The altered architecture of the foot is associated with increased pressure over the ball of the foot, corresponding to the heads of the metatarsals. Also, the normal person constantly shifts the area of pressure in the foot while walking or running, whereas the diabetic patient with neuropathy is unable to do so because of lack of sensory input from the soles. This constant pressure causes calluses with increased pressure and ultimately ulceration in the high-pressure areas. The hyperhidrosis, cracked and dry skin, and increased small blood vessel flow due to autonomic dysfunction create an excellent milieu for infection to take hold and proliferate (156). Infection develops after the skin breaks down and, combined with ischemia, can eventually lead to gangrene.

Neuropathic arthropathy, or Charcot joint, occurs in the presence of impaired sensation of pain and proprioception, intact motor power, and repeated minor trauma, and usually is found in feet with normal pulses and warmth. The clinical course may be one of acute, painful joint destruction, but usually is painless and characterized by nonedematous enlargement of the foot so that the foot becomes shorter, wider, everted, and externally rotated with a flattening of the arch. The gait becomes abnormal, and clubfoot develops. It is usually limited to the ankle and tarsal joints in DM. Pathologic radiographic features include osteopenia, bone lysis, fragmentation, and eburnation. There is disarticulation and dissolution of the joints with bony overgrowth followed by calcification in and around the joints. Eventually pressure ulcers, infection, and osteomyelitis develop. The feet of diabetic patients often have bounding pulses that suggest an adequate large blood vessel supply. The impression is erroneous, however, and these pulses are now thought to be due to the shunting of blood through the small arteriovenous fistulas normally regulated by the sympathetic nervous system. The enhanced blood flow may be conducive to excessive bone resorption, fractures, and osteoarthropathy (11,157).

Charcot joint may present acutely with severe pain, a warm to hot foot with increased blood flow (despite decreased warm sensory perception and vibration detection), and clear evidence of acute osteopenia. Pain and inflammation respond to bisphosphonate within 3 to 4 weeks (158). An important factor in the development of Charcot joint is equinovarus deformity due to Achilles tendon shortening. This is correctable by surgical lengthening. Orthotic devices are the only means of treatment once destruction of a Charcot joint is complete.

Emerging evidence supports the notion that the cause of Charcot neuroarthropathy is not simply repeated minor trauma. Animal studies have documented the need for denervation combined with repeated minor trauma to reproduce the findings in people with Charcot joint. Our own findings suggest that there is a failure of restriction of blood flow that normally occurs in the patient with neuropathy (159) (Fig. 91.5), and that this is associated with osteopenia (160). The osteopenia predisposes the small bones of the foot to small fractures with minimal provocation, especially with the development of equinus. Equinus is due to the shortening of the Achilles tendon
as a result of destruction of collagen fiber elasticity, presumably caused by accumulation of advanced glycation end products (161). We postulate the sequence of events illustrated in Figure 91.5.

Once neuropathy is diagnosed, therapy can be instituted with the goals of ameliorating symptoms and, it is hoped, preventing progression. Successful management of these syndromes must be geared to the individual pathogenic processes. Patients with large fiber neuropathies are incoordinate and ataxic. As a result, they are 17 times more likely to fall than their non-neuropathic counterparts (162). Older subjects have more neuropathy than younger subjects, especially that involving large fibers. It is vitally important to do everything possible to improve strength and balance in the person with large fiber neuropathy. It has been demonstrated that high-intensity strength training, including leg presses, knee extension, back extension, lateral pull-downs, and abdominal extension, in older people increases the strength of a variety of muscles, thereby reducing risk factors for osteoporotic bone fractures. More important, strength training results in improved coordination and balance, quantifiable with backward tandem walking (163). Thus, to preserve nerve function in older people, it is vital to embark on a program of strength training and improvement of balance. Low-impact activities such as Tai Chi are particularly helpful.

To reduce the likelihood of osteopenia contributing to further joint destruction, we now treat all such patients at risk for joint destruction, when they are in the inflammatory phase of the condition, with the bisphosphonate pamidronate, given by slow intravenous infusion over 12 hours. This has been reported by Selby and colleagues (158) to be effective in a small group of patients. What is needed, however, is a double-blinded, placebo-controlled study to prove the efficacy of this agent. It is unsafe to treat diabetic patients with the oral forms of these compounds because esophageal dysfunction increases the risk for obstruction and perforation. We now regularly quantitate the pressure in the feet using an F-scan device and the degree of limitation of ankle flexion with goniometry. If flexion is restricted with shortening of the Achilles tendon and distortion of the midfoot, we simply cut the Achilles tendon and resuture it so that the ball of the foot is no longer receiving the full brunt of body weight in contact with the floor. It is also important to ensure that the patient has proper shoes, has orthotics for correction of improper foot alignment, and, if necessary, has the foot reconstructed by someone experienced in this type of problem.

Diabetic autonomic neuropathy may involve any system in the body. Its manifestations are protean, and the onset often is insidious. Using cardiovascular reflex tests, the prevalence is reported to be 17% to 40% (15,171,172,173,174). Of teenagers with type 1 DM, 31% have abnormal test results (15). The relationship with sensorimotor neuropathy is variable, but autonomic and sensory motor abnormalities usually coexist. Of people with peripheral neuropathy, 50% have asymptomatic autonomic neuropathy. When symptoms of autonomic neuropathy are present, the anticipated mortality rate is 15% to 40% within 5 years (50,175,176,177). With gastroparesis, 35% die within 3 years, usually of aspiration pneumonia. Reduced exercise tolerance, edema, paradoxic supine or nocturnal hypertension (178), and intolerance to heat due to defective thermoregulation are consequences of autonomic neuropathy. Silent myocardial infarction, respiratory failure, and sudden death are hazards for the diabetic patient with cardiac autonomic neuropathy (179,180,181,182). It is therefore vitally important to make this diagnosis early so that appropriate intervention can be instituted. For a complete discussion of the clinical and biochemical features of autonomic neuropathy and its management, the reader is referred to our extensive review (183).

A thorough clinical examination is essential in the evaluation of patients suspected of neuropathy, with special attention to the feet, examining for (a) dryness, shiny skin, cracking of the skin; (b) ulceration; (c) loss of hair; (d) levels of loss of sensory modalities, with particular emphasis on vibratory and thermal sensation; (e) reflexes; and (f) motor power. Diabetic neuropathy is diagnosed by exclusion of a variety of other causes of neuropathy (Table 91.5).

Systematic questioning, including a family history of nondiabetic peripheral nerve disease and the presence of toxic, metabolic, mechanical, and vascular causes of nerve disease, should be conducted. If any other potentially neuropathic factors are present, other diagnostic methods should be used to determine the cause of nerve disease. Appropriate laboratory screening for the differential diagnoses (Table 91.3) should be performed.

To aid in the assessment of neuropathy during a clinical examination, coded scores of neurologic function have been developed that allow assignment of broad categories of functional abnormalities based on the clinical examination (e.g., 1 = normal, 2 = mild abnormality, 3 = moderate abnormality, 4 = severe abnormality, and 5 = total loss of function). Both signs and symptoms can be scored, and the nerve symptom score can be maintained as a record of the patient’s response to treatment. Equally important, and neglected in the routine physical examination, are the functional correlates of nerve impairment. The nerve disability score is a useful instrument, and we have developed an activity of daily living instrument to measure the functional impact of neuropathy that ascertains the degree of restriction associated with tasks of living (e.g., the ability to put
on a shirt, button a shirt, squeeze toothpaste, use a fork, turn the pages of a book). We also routinely have the patient fill out a neuropathy quality-of-life questionnaire, which, in these days of outcomes consciousness, is often the primary interest of the patient and the health-care providers. These tests have not been applied widely in the evaluation of diabetic neuropathy, but they offer a means of evaluating the degree of restriction of activities that are of vital importance to the patient as well as monitoring progress of the neuropathy.

More objective indices are found in the QST (156) and QAFT (178). Combined, these tests cover vibratory, proprioceptive, tactile, pain, thermal, and autonomic function (37,185,186). QST provides standardized procedures for evaluating neuropathy that are sensitive, specific, and reproducible in detecting dysfunction before symptoms appear (183). We use standardized equipment to determine levels of cutaneous sensitivity. This equipment provides (a) interval data, (b) more precise methods of stimulus presentation, and (c) minimal subject and operator bias.

The QAFT consists of a series of simple, noninvasive tests for detecting cardiovascular autonomic neuropathy. Developed by Ewing and Clarke (187), they have been successfully applied by many clinicians (188,189,190,191). These tests can be performed at the bedside and provide an index of the neuropathy that correlates roughly with the vibration threshold and certain other measures of somatic and gut neuropathy (192,193). The precise relationship between autonomic and somatic neuropathy remains unclear. Positive correlations of various degrees using different indices of somatic function have been reported (184,192,193,194,195), but further study of the relationship between autonomic and somatic large and small fiber involvement in diabetic neuropathy is required.

Laser Doppler measurement of cutaneous blood flow in the pulp of the fingers and toes may be the most sensitive measurement of sympathetic autonomic dysfunction in the extremities (196). Flow autoregulation in these areas is modulated by sympathetic input and hence depends on the integrity of those fibers, providing another measure of autonomic function.

Quantitative sensory testing is the determination of the sensory threshold, defined as the minimal energy reliably detected for a particular modality. It is a logical extension of the sensory portion of the clinical neurologic examination and has the principal advantage of assigning a numeric value. A number of relatively inexpensive devices have been developed that allow suitable assessment of somatosensory function, including vibration, thermal energy, and light touch.

Quantitative sensory testing can be used to document subtle sensory loss, characterize patients at the onset of a clinical trial, and monitor a modality known to be associated with a specific complication of diabetic neuropathy (197,198,199). The evaluation contributes to the differentiation of the relative deficit in small-diameter (e.g., temperature) versus large-diameter (e.g., vibration) axons, as well as polyneuropathy versus mononeuropathy.

Calibration and units of measure should be expressed in standard terminology, such as micrometers (vibration), degrees Celsius (temperature), and dynes per square centimeter (light touch), and the physical dimensions of stimulation (e.g., waveform, frequency, rise time) should be reported.

Two testing procedures are emerging as the standards in the field: (a) modified up-down stimulation with two alternative forced-choice responses, and (b) ramping of stimulus intensity (method of limits) combined with a yes-no paradigm (186). With the first procedure, the most common definition of threshold is the stimulus intensity corresponding to the 75% correct response point. With the second procedure, threshold usually is defined as the stimulus intensity corresponding to 50% correct detection. To be acceptable, the two alternative procedures must allow sufficient reversals to converge on a true threshold. The yes-no paradigm must include catch trials (where stimulus is not delivered) in sufficient numbers and variable ramps to minimize guessing. More details concerning the different psychophysical methods, response paradigms, and their critical evaluation can be found in Maurissen’s study (200).

Biopsy of nerve tissue may be helpful for excluding other causes of neuropathy, such as the familial hypertrophic forms, amyloid, sarcoid, and other granulomata, but the pathologic process of diabetic neuropathy is not unique. The pathologic process of early sensorimotor neuropathy is not known, but in established neuropathy, the characteristic picture is that of distal fiber loss and degeneration. Histologic data have been reexamined in light of the subtle asymmetries and focal nature of the clinical presentation; findings have supported the hypothesis that there are differences in the types of neuropathies occurring in type 1 versus type 2 DM. Unlike type 1 DM, in type 2 the major observation is Wallerian degeneration, which may be patchy and irregular, supporting the notion of a vascular origin of the disease in this form of DM (12,92,94,249,250,251,252). Although these findings cannot be distinguished clearly from other causes of neuropathy, a microvascular occlusive picture in the absence of known vasculitides may be pathognomonic of DM. Skin biopsy is receiving increasing recognition for documentation of disease of small nerve fibers. Several elegant studies have shown the loss of axons expressing the protein PGP 9.5 in idiopathic small fiber neuropathies. Studies are under way to determine if this is a feature of the small fiber component of diabetic neuropathy (253).

Figure 91.1 indicates that even with ongoing destructive forces, all is not lost, and therapy can be directed toward regeneration of damaged nerves. We have projected specific growth factors for specific nerve deficits, such as NGF for small fiber, neurotropin 3 for large fibers, and IGF-1 and IGF-2 for motor fibers.

Nerve growth factor is one of several neurotrophic factors that is known to play an important role in the development, maintenance, and survival of neuronal tissues. NGF binds to high- and low-affinity NGF receptors. The high-affinity receptor is a tyrosine kinase transmembrane protein called TrkA; the low-affinity receptor is a 75-kd glycoprotein referred to as p75. Both are present on small, unmyelinated fibers of the sensory neurons of the peripheral nervous system, on sympathetic neurons in the autonomic nervous system, and in regions of the central nervous system. Alterations of the small, unmyelinated DRG neurons are responsible for the pain, paresthesias, and numbness of diabetic neuropathy (278,279,280). There is now considerable evidence in animal models of DM for decreases in the expression of NGF and TrkA, reduced retrograde transport of NGF, and diminished support of small, unmyelinated neurons and their neuropeptide neurotransmitters, substance P and CGRP, both potent vasodilators alluded to previously (281,282,283,284). Furthermore, recombinant NGF (rNGF) administration restores these neuropeptide levels toward normal and prevents the manifestations of sensory neuropathy in animals (285).

We completed a phase II study in humans testing NGF (286). This was a 15-center, double-blinded, placebo-controlled study of the safety and efficacy of rNGF in 250 subjects with symptomatic small fiber neuropathy. NGF decreased the amount of neurologic impairment measurable on clinical examination of the lower limbs and improved small fiber function
measured as the ability to detect cooling (Aδ fibers) and the ability to perceive heat pain (C fibers) (HP; 5.0) compared with placebo. These results are consistent with NGF’s postulated actions on TrkA receptors present on small fiber neurons. Data from two large trials just concluded in the United States and elsewhere do not support our phase II study inasmuch as the objective of a greater response in the NGF-treated groups compared with the placebo-treated patients was not achieved. For this reason, the clinical trials have been stopped, but it is not readily apparent why this discrepancy should have occurred.

Painful neuropathy can actually be precipitated by initiation of diabetic therapy, although it usually subsides with ongoing treatment. The syndrome of acute painful neuropathy associated with marked weight loss (diabetic neuropathic cachexia) also responds to control of hyperglycemia, although symptoms may persist for as long as 6 to 18 months, requiring constant reassurance to the patient. We believe this to be a variant of CIDP, and we have seen dramatic responses to IVIG in a small number of patients.