An Overview of Peripheral Neuropathy in Diabetes

How Prevalent is Neuropathy?

Peripheral neuropathy is the most prevalent complication of type 2 diabetes. The 2004 National Health and Nutrition Examination Survey (NHANES) on lower extremity complications revealed that close to 10% of people with diabetes have peripheral arterial disease (PAD), but close to 30% have neuropathy. Disturbingly, the survey also showed that over 7% have an active foot ulcer, a frequent cause of hospitalization and a common pathway to amputation. All of these figures are likely underestimated, since the NHANES is a community survey and thus does not include hospitalized and infirm patients. All told, the survey showed that close to 40% of people with diabetes have some identifiable lower extremity complications, of which neuropathy is the most common. Furthermore, in a 10-year natural history study of patients with type 2 diabetes from the time of diagnosis, there was an increased prevalence of pain, paresthesias, loss of reflexes, and loss of vibratory sense.¹ These are signs and symptoms of more severe neuropathy, indicating probable loss of protective sensation (LOPS). Overall, approximately, there was a 30% prevalence in patients with diabetes of LOPS, which is the most common risk factor for foot ulceration. The natural history observations at presentation are that close to 10% have peripheral neuropathy, but after 10 years, 40% have definite or probable neuropathy, again making it the most prevalent complication.

Beyond hyperglycemia. The observation that peripheral neuropathy is present at the time of diagnosis implicates the involvement of the pre-diabetic state in pathogenesis. For many years, the focus of our understanding of neuropathy was its relationship to hyperglycemia. Indeed, the Diabetes Control and Complications Trial showed that tight glycemic control reduced the incidence of neuropathy compared to standard treatment in type 1 diabetics. In part, glucose is metabolized by the polyol pathway, which includes aldose reductase, an enzyme that converts it to a non-metabolizable sugar, sorbitol. Hyperglycemia then leads to an accumulation in nerve sorbitol. Nerve ischemia from microvascular disease is also present and leads to increased oxidative stress in the nerve, loss of neurotrophic support, immune stimulation and altered protein synthesis. The postulate was that sorbitol also depleted myoinositol from the membrane, leading to nerve injury and abnormal conduction. Thus, a complex theory on the pathophysiology of diabetic peripheral neuropathy (DPN) stems from a focus on the primary role of hyperglycemia. In a landmark longitudinal study of type 1 diabetes and DPN, Tesfaye et al² found that other risk factors were implicated. All the type 1 subjects had high blood glucose, but it was those with cardiovascular risk factors that developed DPN—they were smokers or had high triglycerides, high cholesterol, and hypertension. The American Diabetes Association (ADA) seeks to prevent diabetes and heart disease, and these are the modifiable and non-modifiable risk factors, including insulin resistance. Many are associated with peripheral neuropathy. Thus, for microvascular complications, the clock starts ticking with the onset of hyperglycemia, that is, retinopathy and nephropathy. However, for macrovascular complications, including neuropathy, the clock starts ticking well before the onset of diabetes, a state characterized by insulin resistance, lipid abnormalities, hypertension, inflammation, and endothelial dysfunction. What some may call “metabolic syndrome” has its primacy in insulin resistance, leading to these clinical manifestations, as well as inflammation and endothelial dysfunction. We can refocus our attention to how cardiovascular risk factors have a causal role in DPN.

Endothelial dysfunction and neuropathy. It is well recognized that the endothelium is biologically active. In type 2 diabetes, there are deranged nitric oxide (NO) pathways. This altered NO response is in part due to the effect of hyperglycemia on the endothelium itself, as glucose transport in that cell is insulin-independent and not down regulated by hyperglycemia. Interestingly, there is also impaired NO response because of insulin resistance and from abnormal free fatty acid production. Thus, defective insulin signaling affects not only glucose transport, but also NO response. Rich and Veves,³ using laser Doppler, looked at patients with neuropathy and endothelial function. The endothelium produces nitric oxide in response to acetylcholine. Normal endothelium responses occurred in controls and diabetic patients, but in patients with diabetic neuropathy, endothelial response was impaired. Subjects with Charcot neuroarthropathy or with vascular disease demonstrated an impaired endothelial response, but that is attributable to the associated presence of DPN. These patients also exhibit an impaired response to nitroprusside, an endothelium-independent vasodilator. An impaired response here implies a problem with the vessel wall in neuropathy, Charcot and vascular disease. Endothelial dysfunction in neuropathy can be caused by several things, including insulin resistance, metabolic abnormalities and advanced glycation end products. It results in decreased capillary flow, nerve hypoxia, and then nerve dysfunction. Endothelial dysfunction seems to have primacy in diabetic neuropathy. Thus, we are moving away from blood sugar as a primary cause into more vascular- and cardiovascular-related causes.

Classification

The ADA recently published an expert consensus for the classification and treatment of neuropathies. There are generalized symmetric neuropathies in diabetes and then there are focal and multi-focal neuropathies. We will be discussing primarily the chronic sensorimotor, generalized symmetric neuropathy, but will look at autonomic and proximal neuropathy, which used to be called amyotrophy, and also co-existing chronic inflammatory demyelinating polyneuropathy (CIDP). Unusual neuropathies. Diabetic proximal neuropathies are also uncommon and difficult to characterize. Pascoe et al⁴ compiled the different findings and treatment responses of proximal neuropathies because they are a mixed group. Mayo Clinic has done the most work in proximal neuropathies, which are a heterogeneous group and are not as common; patients present with weakness. It could be amyotrophy, which is associated with severe weakness and weight loss. It is also associated with the initiation of insulin therapy. Chronic inflammatory demyelinating poly neuropathies are seen in this group. The key is that they are demyelinating. Diabetes itself is usually just axonal, not demyelinating. Since there probably is an autoimmune component, some respond to intravenous immunoglobulin as an immune modulator. There is differential diagnosis for proximal neuropathies. It could be vasculitis, chronic inflammatory demyelinating polyneuropathy, amonoclonal gammopathy, such as a multiple myeloma, and then diabetes. Autonomic neuropathy is probably the largest under- appreciated area in the field of neuropathy. Patients can present with abnormal swelling also affecting the feet, less sweat, less oil, drier skin; also associated is impotence and heart rate abnormalities. Regarding the cardiovascular system, cardiac autonomic neuropathy (CAN) has been shown to be the most important risk factor for silent ischemia in our patients with diabetes. CAN is presently diagnosed with R-R interval when there is absent or reduced variation with respiration or with Valsalva maneuver. In the ADA statement, it is underscored as a risk factor for silent ischemia and sudden death.

Chronic sensorimotor neuropathy. As we noted, sensory motor neuropathy affects about 40% of diabetics. Its symptoms are progressive. It usually involves the longest nerves first and then marches back inexorably as the years pass. The main symptom is numbness, but the most important sign is a LOPS. Sensorimotor neuropathy can be painful, and is characterized by reduced reflexes and muscle atrophy. When there is motor involvement, foot deformities can occur which result in increased pressure. Increased pressure in patients with LOPS can easily lead to foot ulcers. The neuropathic foot is associated with a great deal of muscle atrophy (intrinsic minus) and loss of fat, resulting in deformities and increased plantar pressures. The prominent veins are visible due to the increased blood flow. Neuropathy gives more blood flow at rest than a non-neuropathic patient. There is the beginning of rigid, bony deformity and a rigid foot. Diagnosis is based on symptoms and signs. Electrophysiology and nerve biopsies are rarely needed for diagnosis, but are standard endpoints in clinical trial design. A major workup is not necessary to identify these patients. The simplest method is use of the 10-g (5.07 Semmes-Weinstein) sensory testing nylon filament. Press on the foot and allow it to buckle. If the patient cannot feel it, they have loss of protective sensation and peripheral neuropathy, and are at risk for foot ulceration. This test can also be done with vibration. A vibratory perception threshold is more quantitative. The gain is turned up until the patient registers the vibration—greater than 25 mV would be diagnostic of loss of protective sensation. A neuropathy disability score gives you a composite, involving a pinprick, light touch, hot and cold, vibration, and ankle reflexes. It is used commonly in neuropathy research studies. The Semmes-Weinstein is more than adequate for most clinics. Sometimes it is good to back it up with the tuning fork vibration as well. For those that want to track the progression of the neuropathy from year to year, we recommend a vibratory perception threshold, because a quantitative number is available. The neuropathy disability score, the vibratory perception threshold, and the Semmes-Weinstein monofilament all tie, meaning with any of those modalities, a diagnosis can be made.

What about a differential diagnosis? Alcohol abuse would be common. B12 deficiency is also a possibility, especially if the patient has autoimmunity and is a type 1 diabetic. Generally, if the patient has diabetes, and has symmetric neuropathy and a loss of protective sensation, then a diagnosis of DPN can be made. A referral to a neurologist is generally not necessary.

How Do We Prevent or Slow the Progression of Neuropathy?

Tight glycemic control is the only thing that has been shown to have any neuroprotection. As mentioned, it has been best shown in type 1 diabetics. For type 2, it is less clear, because in type 2 the cardiovascular risk factors are independent of the glycemic control. Most research has been done on the polyol pathway, thinking that if we could inhibit the aldose reductase and prevent the accumulation of sorbitol it should protect the nerve. Yet the idea that sorbitol is the villain has not been borne out after a great deal of expense and time. The final, common path may be through endothelial dysfunction and nerve ischemia, but more research is needed. Studies have been done in the last two decades, but none of these drugs made it through to Phase 3 with the US Food and Drug Administration (FDA). Aldose reductase inhibitors are ineffective. Most studies looking at alpha lipoic acid have been negative, but a few have suggested an improvement. Alpha lipoic acid is not harmful and patients can get it at a health store, so there are no strong objections to its use. In a major study from Genentech looking at nerve growth factor, the placebo did better than the treated group. Gamma linolenic acid has been shown to be ineffective. ACE inhibitors have been shown to be effective in one trial by Malik, which supports the idea that if we can improve endothelial function and nerve blood-flow, it would be effective. A couple of early phase one trials are looking at the VEGF gene and the VEGF Zinc Finger Protein to try to stimulate angiogenesis to the nerve. In animal studies, preclinical VEGF has shown a neuroprotective effect. Ruboxistaurin is a protein kinase C–beta isoform inhibitor. It did not seem to be effective and it is no longer in trial. Benfotiamine also improves endothelial function. It is a thiamine derivative and seems to help with complications of diabetes in general. In an animal study, it did show effectiveness. Pyridoxamine is a B-6 analog that prevents the formation of advanced glycation end products and it has also been shown to be effective in an animal model. While we have not yet mastered neuroprotection, there are some interesting ongoing studies. Our group is currently involved with a study using Metanx, a medical food by Pamlab, that is a composite of active metabolites of folic acid, B6, and B12; L-methylfolate, methylcobalamine, pyridoxal 5’-phosphate. It has been shown to improve endothelial function as measured by FMD. Through improving endothelial function, it may have a benefit as a neuroprotective agent. The vibratory procession threshold at 6 months will be the endpoint.

Treating Neuropathic Pain

Treatment options do exist for neuropathic pain. The commonly used visual analog scale (VAS) values range from 1 to 10: How bad is your pain? In pain trials, the VAS score is 6 and the best result is a reduction to approximately 4. None of these drugs eliminate the pain completely; they just make it more tolerable. The FDA has approved two drugs for neuropathic pain. One is the tricyclic duloxetine (Cymbalta) and the other one is the anticonvulsant pregabalin (Lyrica).

Tricyclics. The best evidence for pain relief of neuropathy is with tricyclics. Amitriptyline (Elavil, Endep) does work, although there are tolerability issues. In a 1992 study on tricyclics^,4 amitriptyline did well, as did desipramine (Norpramin, Pertofrane). Note that selective serotonin reuptake inhibitors (SSRIs) have not been shown to be effective. Fluoxetine (Prozac) did not do much better than placebo.

Serotonin-norepinephrine reuptake inhibitors (SNRIs). Duloxetine (Cymbalta) is the first serotonin norepinephrine inhibitor approved by the FDA for neuropathic pain. Duloxetine’s phase 2 trial improved pain by as much as three points on the scale. Patients with the higher dose of duloxetine required less additional analgesics.

Anticonvulsants. As a class, the anticonvulsant drugs most likely work as a class, although right now only pregabalin is approved. Gabapentin (Neurontin) has a positive safety profile except for the potential side effects of dizziness and somnolence. The FDA never approved it for neuropathic pain; nonetheless, one off-label study showed reduction in pain. Gabapentin is commonly used and a less expensive generic version exists, but the problem with gabapentin is the titration. The next generation after gabapentin is pregabalin. Pregabalin works on the alpha 2-delta subunit of the calcium channels. It seems only to work on excited nerves. The alpha 2-delta subunit is over-expressing, what is called excited nerves or hyperexcited neurons. Pregabalin seems to be a membrane stabilizer, which is probably how it reduces neurotransmitter release and helps with pain. The advantage of pregabalin over gabapentin is its linear pharmacokinetic profile, making it very easy to titrate. In the pregabalin study by Rosenstock et al,⁵ the pain score again went from 6 to 4, with almost identical patient profiles. Over 40% of patients had a greater than 50% reduction in pain. Nobody gets a 100% reduction in pain; do not have that expectation for yourself or your patients. A 50% reduction in pain and more sleep at night for your patients may be the best that can be done. If an opiate must be added, there is some evidence from a small study that tramadol (Ultram) may be particularly good and it does not seem to have much addictive potential. Carbamazepine (Tegretol) has been used in the past. Clinically, this seems to help, especially with the lacerating, electric shock-type pain many patients get at night. There is some literature on topiramate (Topamax) and lamotrigine (Lamictal).

Conclusion

The ADA expert consensus offers a treatment algorithm for DPN. The first thing to do is to make the diagnosis and exclude non-diabetic causes. Next, stabilize glycemic control. Boulton et al⁷ showed that swings in glycemia make neuropathic pain worse, so if the glycemic control can be stabilized, it may help. If pharmacologic therapy is used, the first choice should be a tricyclic, because it is inexpensive and efficacious. This is according to the ADA consensus; however, there are tolerability issues experienced by some patients, largely CNS, such as alertness, somnolence, and dizziness. After a tricyclic, the next option would be an antidepressant or an anticonvulsant such as duloxetine or pregabalin. If that doesn’t work, an opiate can be added safely. Tramadol has been shown to be effective in treating painful neuropathy. If an opiate proves ineffective, the next step would be a referral for pain management. Remember that drugs that improve endothelial function may help protect the nerve. Thus reducing all cardiovascular risk factors such as smoking, lipids, and hypertension, which will help in reducing major cardiovascular events, may also help slow the progression of DPN.

Disclosure: Dr. Sheehan has disclosed that he has received grant/research support from Tissue Repair Company, PamLab, Genzyme and Sanofi-Aventis. He is the Director of Greystone Pharmaceuticals. He is a member of the scientific advisory board for Advanced bioHealing, a consultant for Heal Or, Hypermed, and Calretex. He is also a member of the speakers’ bureau for BMS/Sanofi, Merck, and Organogenesis.

Correspondence: Peter Sheehan, MD, President, Sheehan Health Management Corp., 422 East 72nd Street, Apartment 24 E, New York, NY 10021; E-mail: psheehan@verizon.net.