The mechanism of RICF likely relates to the unique metabolic characteristics described previously. Diabetic nerves have a severalfold increase in nerve energy substrates (glucose, fructose, glycogen) (11). Their energy requirements are sufficiently low to permit them to continue to conduct impulses on adenosine triphosphate (ATP) derived from anaerobic metabolism (1). Their energy requirements may also be downregulated. The relatively large endoneurial space (relative to brain) results in less susceptibility to lactate- and potassium-induced conduction failure (11).

The precise pathophysiology is not known. Circumstantial evidence suggests that the combination of ischemia and mechanical susceptibility might be responsible. Evidence for microvascular ischemia clinically includes the acute or subacute onset of, for example, third-nerve palsy in vasculopathic older patients. Pathologic evidence is limited but includes the presence of microinfarcts, multifocal pathology, and fibrinoid change (16). Such changes have been described in two diabetic patients; one died shortly after the onset of oculomotor palsy (17); in the second, a 73-year-old man with diabetes who died shortly after the onset of a focal limb neuropathy, the obturator, femoral, sciatic, and posterior tibial nerves were examined (18). Combined changes of florid round cell infiltration and microvascular changes have been interpreted as being due to a combination of microvascular ischemia and immune-mediated mechanisms. Three sets of findings are consistently found in chronic diabetic neuropathy: first, a thickening of basal lamina (19); second,
endothelial-cell and smooth-muscle proliferation (20,21) with increased variance of capillary caliber and the closing of some capillaries (20); third, the tendency of fiber loss to be multifocal rather than diffuse (22,23). The combination of microvascular ischemia, connective tissue changes (24), and myelin alterations could presumably render the nerve trunk susceptible to compression injury.

Vasoregulation of nerve microvessels is regulated at the arteriolar, venular, and endothelial cell levels. In the diabetic state, endothelial changes include those of AGE-RAGE-NF-κB [advanced glycosylation end products (AGE)–AGE receptor- nuclear factor κB (NF-κB)], cell adhesion molecules, nitric oxide, and eicosanoids. Experimental diabetic neuropathy results in a reduction in nerve blood flow of about 50% (33,36), which begins by the first week of diabetes (36) and affects the cell body of sensory and sympathetic neurons as well as the axon (37). In diabetes, the combination of increased whole-blood or plasma viscosity, increased aggregability, and reduced red blood cell deformability results in a reduction of blood flow and stagnation hypoxia (1).

Regional nerve blood flow is regulated by systemic blood pressure (4) and the balance between neural vasoconstrictors and vasodilators. Vasoconstriction is known to be mediated by epineurial α-adrenergic (38,39), vasopressin (40), and endothelin receptors (41). The 50% effective concentration (EC₅₀) of endothelin, using identical methodology, is about three orders of magnitude lower, indicating greater potency, than that of adrenoreceptors (41). The EC₅₀ of vasopressin is of an order similar to that of endothelin. These potent vasoconstrictor actions are balanced by vasodilatation and are mediated by calcitonin gene-related peptide (42), substance P (42), nitric oxide (NO) (43), and the prostaglandins (44).

Alterations in microvessels in experimental diabetic neuropathy have been reported. These are reported to be excessively sensitive to α-adrenergic agonists (45) and endothelin (46,47). The nerve conduction and blood-flow deficits are prevented by specific antagonists to endothelin (48) and adrenoreceptors (49). Deficits in vasodilators are supported by restoration of blood flow and nerve conduction by an NO donor (50) and vasodilator prostaglandins (51). The reduction of NO is likely the most important in its magnitude and implications (43). The polyol pathway is overactive and competes with nitric oxide synthase for NADPH (reduced form of nicotinamide adenine dinucleotide phosphate). By siphoning off NADPH, chronic hyperglycemia results in a deficiency of NO in nerve endothelium (43,50). Endoneurial NO deficiency has a number of important consequences in the microvasculature. It normally inhibits the expression of P-selectin, intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1), and other adhesion molecules (52), a mechanism that is mediated
through inhibition of protein kinase C and by preventing the activation of NF-κB (53). NO also inhibits the cytoassembly of NADPH oxidase (54), thereby attenuating the release of superoxide by leukocytes (55). Inhibition of NO, therefore, explains a number of changes seen in experimental diabetic neuropathy (expression of adhesion molecules, activation of NF-κB, expression of cytokines, and generation of superoxide anions).

Three main pathologic alterations affect human diabetic peripheral nerve. First, thickening of endoneurial capillary wall and thickening and reduplication of basal lamina have consistently been found (56,57), especially in more recent studies (16,58). Basal lamina thickening, although characteristic, is not specific to the diabetic state, occurring especially with aging and a number of other neuropathies. Second, alterations in capillaries are consistently present. These changes include a reduction in capillary density, reduced capillary area, and capillary closure (16,20,58). Third, the pattern of fiber loss is multifocal, suggestive of ischemic fiber loss (22,23). In a three-dimensional reconstruction of the distribution of fiber loss using autopsy specimens from lumbosacral trunk to sural nerve biopsy specimens, the changes seen in sural nerve manifest the cumulative loss of fibers, whereas multifocal fiber loss is better appreciated in more proximal nerve trunks (59). Although these alterations have been well defined, their specificity has been questioned (60). These studies are all important in attempting to define morphometric abnormalities of endoneurial vessels and fibers. They suffer from their ability to define the results but inability to define the pathogenetic process.

The most concerted and coherent studies of perfusion and oxygenation of human diabetic nerves have been undertaken by the Sheffield group. They were the first to demonstrate reduced oxygen tension in human diabetic neuropathy. They measured endoneurial oxygen tension in exposed sural nerve of 11 patients with diabetic neuropathy (34). Correlative pathologic studies in a separate study were supportive, showing reduced endoneurial capillary density and changes in capillary basement membrane and endothelial cells and in the derived total diffusion barrier that was significantly impaired in the nerves of patients with severe diabetic neuropathy as compared with control subjects (58). Blood flow and oxygen tension recordings in the operating room are not trivial. Recordings using laser Doppler methodology have been reported to show no change or even a slight increase in blood flow and oxygen tension (61). These results are un-interpretable in diabetic neuropathy, because it measures total and not nutritive flow, where shunt flow may actually be increased (62).

A number of innovative approaches have been developed to measure flow and oxygenation. Nerve photography and fluorescein angiography demonstrate delayed fluorescein appearance time and intensity of fluorescence in human diabetic neuropathy (62). These investigators observed direct epineurial arteriovenous shunting in the majority of diabetic neuropathic patients. A novel approach using microlightguide spectrophotometry to measure intravascular oxygen saturation and blood flow in human sural nerve after an intravenous injection of sodium fluorescein has confirmed impaired nerve blood flow and tissue deoxygenation. There was a correlation between rise time, nerve oxygen saturation, glycemic control, and sural nerve sensory conduction velocity (63).

Angiotensin-converting enzyme (ACE) inhibition with lisinopril administered to 13 hypertensive patients with diabetic neuropathy was reported to improve nerve conduction indices and sensory perception thresholds (64). A small, randomized, double-blind, placebo-controlled trial on the effect of the ACE inhibitor trandolapril on 41 normotensive patients with type 1 or type 2 diabetes and mild neuropathy has appeared (65). Changes in the neuropathy symptom and deficit scores, vibration-perception threshold, peripheral-nerve electrophysiology, and cardiovascular autonomic function were assessed at 6 and 12 months. The primary end point was the change in peroneal nerve motor conduction velocity. Peroneal motor nerve conduction velocity (p = 0.03) and M-wave amplitude (p = 0.03) increased, the F-wave latency (p = 0.03) decreased, and sural nerve action potential amplitude increased (p = 0.04) significantly after 12 months of treatment with trandolapril as compared with response to placebo.

The study is interesting and has potentially important implications. There is controversy over whether ischemia is a central mechanism or is one of many mechanisms in the production of diabetic neuropathy. In view of the strength of the evidence of the benefits of ACE inhibition on renal function, it might be difficult to set up a specific study on neuropathy, in contrast to a neuropathy piggyback study similar to the Diabetes Control and Complications Trial.

Oxidative stress refers to a shift in the pro-oxidant:antioxidant status toward the former. Plasma levels of lipid peroxide are increased in human diabetes (69,70,71,72,73,74) and occur early, even in young persons with diabetes (75). The highest levels were reported in patients with microvascular angiopathy, as manifested as retinopathy or microalbuminuria; lowest in those patients with diabetes without angiopathy (69,76); and normal in patients with well-controlled diabetes (69). The relationship may relate to the observation that low-density lipoproteins of patients with diabetes are significantly more oxidizable than are those of controls, an abnormality that is correctable by 6 weeks of treatment with the antioxidant probucol (77). Presumably, patients with diabetes with angiopathy who have higher levels of low-density lipoproteins will have the greatest lipid peroxidation.

More recent data are more consistent with increased oxidant stress being secondary to metabolic status rather than to complications (78), including a positive correlation with glycosylated hemoglobin and a negative correlation of malondialdehyde with glutathione peroxidase (79). Recent work has focused on combining measurements of lipid peroxidation (increased)
with indices of total antioxidant defense (reduced) in diabetes (80,81). Additional emphasis has been placed on evaluation of indices that reflect specific mechanisms and tissues affected by oxidant stress. Measurements include those of increased activation of the oxidative stress-sensitive transcription factor NF-κB (82). Greater importance also is placed on products that specifically reflect neural or DNA damage, such as 4-hydroxynonenal and 8-oxo, 2′-deoxyguanosine in relevant tissue (83). Urinary 8-oxo, 2′-deoxyguanosine excretion and the 8-oxo, 2′-deoxyguanosine content in the mononuclear cells from diabetic patients with complications were higher than those from diabetic patients without complications (84,85).

Glutathione (GSH) is reduced in erythrocytes from patients with type 2 diabetes, with a corresponding increase in oxidized glutathione (GSSG) (86,87). In subsequent studies, this same group additionally demonstrated reductions in the glutathione synthesizing enzyme, γ-glutamylcysteine synthetase, and the transport of thiol [S-(s,4-dintrophenyl)glutathione] in erythrocytes of these patients. These abnormalities were reversible with improved glycemic control. They also demonstrated that a medium high in glucose augments the toxicity of xenobiotics on K562 cells, associated with reduction in both the enzyme and its mRNA (88). Lipid peroxidation is increased and antioxidant enzymes are reduced in the erythrocytes of persons with type 2 diabetes. Changes in GSH and antioxidant enzymes occur early, even at onset of type 1 diabetes (75).

Erythrocyte cuprozinc superoxide dismutase (SOD) is reduced in patients with type 2 diabetes (70,89), and catalase is variably reduced (89). It is suggested that this reduction is mediated by the accumulation of intracellular H₂O₂ (90). Changes in SOD occur early (75).