Nerve Decompression in the Management of
Diabetic Peripheral Neuropathy: Is There Evidence?

If we can surgically decompress patients with active entrapment,
we may be able to alleviate their pain.


As a young surgeon, I treated patients with carpal tunnel syndrome, and many of these patients had diabetes. They would ask, “Can you help my feet?” I confessed at the time that I did not know that much about feet.

How does what we see in the hand relate to what we see in the foot? People have numbness in their fingers and hands, which bothers them at night and they feel like their whole hands are numb. They get clumsy and they drop things. This numbness is very similar to what people with diabetic peripheral neuropathy (DPN) complain about in reference to their feet.

The randomized controlled trials that have been done and are ongoing have not given us a real solution for DPN. How can I help these people with painful symptoms?

We know that we must identify patients with DPN early using methods such as calibrated tuning forks and biothesiometers. These tests to evaluate the peripheral nervous system have been in practice for many years, however, they use estimates and algorithms. There are many criticisms of these classical tests. For example, as surgeons we want to distinguish the perineal nerves from the tibial nerve. You cannot do this with vibration detection threshold because it stimulates both nerves.

The two-point discrimination tells us about intervention density, and in 1985 we created a device to make it easier to test. Later, with an aerospace engineer to help evaluate the outcomes from our surgery, we developed a new device: the pressure-specified sensory device. In an early article, we showed that diabetics with ulcerations had poorer sensory results than diabetics without ulcerations.

When we look at the authorities on DPN, this sums it up and has for many years: We do not actually have anything that we can do for the patient. So, I applied the concept of decompressing multiple peripheral nerves in the extremity of a patient. Some patients can wake up in the recovery room and feel some sort of improvement. This helped us refine our hypothesis: Symptoms of DPN — in some patients — are due to compression of peripheral nerves. If this is true, the corollary will be that the symptoms of DPN in some patients are reversible by decompressing of those peripheral nerves.

In 1992 we published a study of 60 diabetic patients. The included cohort had to have a positive Tinel’s sign to signify the superimposed entrapments and no ulcer or amputation. In this group, 8% were said to be normal and 11% had localized entrapment, with the rest having diffuse peripheral neuropathy. Some patients were caught early enough that they were identified with the traditional testing of superimposed nerve compressions. We operated on these people, decompressing a total of 154 different nerves, and they were followed for an average of 30 months. In the total cohort, 88% were improved. In the group that had localized compression, 100% of them were healed completely. The group that had neuropathy with superimposed compressions had 80% improvement and in the group that had diffuse neuropathy, 50% were better. Criticisms? I evaluated my own patients, and I am clearly biased. Also, maybe there was a placebo effect.

In the laboratory, we looked at the basic science work. We studied the pathophysiology of nerve compression in rats and monkeys using a tube that was relatively long with respect of the diameter of the nerve. After 6 months of compression, we saw thinning of the myelin. By 1 year, we have continued thinning and dropout of large mylenated fibers. We can surgically decompress this nerve and regain fibers, although the myelin does not recover its previous thickness.

One important implication when it comes to evaluating our patients is that over time, their symptoms are going to change. Why would diabetes predispose the nerve to chronic compression? First, the nerve is swollen. The nerve swells in a tight anatomic area, and we can postulate decreased blood flow and chronic compression. There is decreased ectoplasmic flow, and this is a slow component that means that building blocks are delayed and the nerve has to repair itself distally. We applied our chronic decompression model and showed that a diabetic banded rat was more likely to have nerve compression and be worse than the nonbanded rat.

Our hypothesis is that the symptoms are related to nerve compression. Prospective studies from what you might term a loose meta-analysis of the 14 published studies, all with the same inclusion criteria, operative technique and outcome measures, have shown good results. We have a cohort of 833 patients with this degree of magnitude of improvement and no new ulcers or amputations.

We have trained 260 surgeons in 42 states in the United States, and 36 of these have begun entering patients into a prospective multicenter study available on the Internet at neuropathyregistry.com. So far, there are 549 patients and 40% have had their second side done. We have seen no new fractures in this group of patients.

How do we select the patient with superimposed compression? We use a positive Tinel’s sign, as simple as that sign is. In this group of patients, remember before we had 69% improvement. Now, we have a positive Tinel’s sign being a positive predictive value of 92%.

If we can decompress patients with active entrapment — and there are many patients with this condition — we may be able to improve their symptoms.

A. Lee Dellon, MD is the founder of The Dellon Institutes for Peripheral Nerve Surgery. He is a plastic surgeon as well as a professor of plastic surgery and neurosurgery at The Johns Hopkins University School of Medicine, the University of Maryland and at the University of Arizona. He specializes in the treatment of diabetic neuropathy as well as other painful peripheral nerve disorders and has trained many surgeons worldwide in the procedures he has developed to relieve pain. He may be reached at 877-DELLON1.

There is clearly a place for surgery in the treatment of common or “garden variety” diabetic neuropathy … it is just that I have not been able to find it.


The real issue with regard to nerve compression and the diabetic patient is should we be doing decompression surgery for diabetic peripheral neuropathy (DPN) as opposed to surgery for entrapment syndromes? While Dr. Dellon’s presentation may have been very eloquent, we have to ask whether the argument can be translated into care of the patient with diabetes.

The clinical presentation of DPN is that it involves small and large somatic and autonomic nerve fibers with an increased incidence (30%) of entrapment. According to the recent position statement of the American Diabetes Association, to make the diagnosis of DPN there must be more than two abnormalities, such as on neurological examination, nerve conduction study, quantitative autonomic function, quantitative sensory or quantitative motor function. The diagnosis of entrapment requires electrophysiologic evidence of a conduction block at the site.

Many people believe diagnosis is a relatively simple task and that we can all do this. But, let us look at the mistakes we have all made with this diagnosis. With mild neuropathy, in which there is a potential for reversal of the situation, the diagnosis is missed 65% to 70% of the time. We still need to improve our ability to recognize the symptoms. We mistake osteoarthritis, Charcot neuroarthropathy, diabetic neuropathy, plantar fasciitis, tarsal tunnel syndrome and fibromyalgia as DPN. Failed clinical trials for DPN can be ascribed to this negligence to define the cause of the pain.

How do we measure outcomes? The symptoms must improve to a significant degree, and it must not be due to worsening of neuropathy. We need to be sure we are not creating a deterioration of nerve function. The change, if possible, should correlate with evidence of improved nerve function such as nerve conduction, intraepidermal nerve fiber, quantitative sensory and quantitative autonomic function tests. The change should be biologically relevant, based upon the view that this is what patients interpret as clinically meaningful. The treatment being evaluated should have a low number needed to treat, but more importantly a high number needed to harm (primum non nocere)!

Objective measures of responsiveness. There are different scoring systems for evaluating neuropathy, eg Dyck’s Neuropathy Symptom Score (NSS), Neuropathy Disability Score (NIS) and the Neuropathy Impairment Score of the Lower Limbs (NISLL). You can also use a simpler neuropathy disability score such as Eva Feldman’s Michigan Neuropathy Scoring System. We should attempt to relate the clinical score to the nerve function. In addition, quantitative measures for threshold of vibration, thermal perception, heat pain, cold pain, touch pressure and nerve conduction can be obtained.

We developed receiver operating characteristics measuring the area under the curves for sensitivity and specificity of small and large fiber function tests. The vibration detection threshold gives about 90% sensitivity if the specificity is held at >90%. It is important to remember that different measures measure very different things, and that the test must be validated for snesitivity, specificity, reproducibility, accuracy and to be able to show a change with therapy. Two-point discrimination has not met these criteria.

Can we tell the difference between entrapment versus DPN or entrapment superimposed on DPN? My colleagues and I have looked at the frequency of entrapment of the tibial nerve in 4,600 patients. If the latency of the perinneal and the tibial nerves are equal, then the diagnosis is DPN. If there is a discrepancy on either side, then there may be entrapment. We found that carpal tunnel syndrome was very common, and that evidence of entrapment of the tibial nerve superimposed on DPN was extremely rare.

In DPN, one sees edema of the canal with atrophy of the small muscles of the feet. What comes first? One of the major causes of DPN is microvascular insufficiency in which damage to the blood vessels by impaired function or occlusion decreases blood supply to both myelinated and unmyelinated nerve fibers. Can one recognize the patient with compression in the tunnel in the face of DPN? This may be extremely difficult.

Do we know its entrapment? Let’s look at the clinical features of 1,528 hands with a neurological diagnosis of carpal tunnel syndrome; symptoms were not restricted to the hand in 31% of cases. Conclusions were wrong 40% of the time, pain was not as reliable as paraesthesiae and Tinel’s sign was positive only 34% of the time. False positives in this group occurred in patients with fibromyalgia and tendonitis. In fibromyalgia there is a large proliferation of the small intraepidermal nerve fibers — the opposite of what occurs in DPN.

Is improvement due to the intervention? What about the placebo effect? Why do you see improvement in people when you operate on them, and they come out one day later and tell you everything is better? Are there ways to cure people using nonsurgical methods such as stochastic resonance, magnetic therapy and infrared therapy? You can enhance sensation in the diabetic neuropathic foot with mechanical noise or vibration prior to testing. In these studies, however, there is no real measure of improvement, but simply a measure of changing the detection of a modality.

Dr. Dellon, however, found a 20% failure of the predictive value of a Tinel’s sign. When I touch, vibrate or add pressure to one foot, the threshold in the other limb changes. So, now when I am trying to do this, if I stimulate the other limb, I will produce a reduction in the patient’s ability to perceive. Now, if I stimulate the same limb, I will improve the patient’s ability and if I stimulate the same limb further with noise plus measuring two-point discrimination, I will now give the patient a greatly enhanced two-point discrimination. So, this says to me, can I really trust this as a measure of what I am looking at, when I can do this with noise?

Could we possibly do damage with surgery? If you have good nerve function, you have pain. But, if you damage the system when you have pain and are through the pain threshold, pain can also disappear. We do not want to take people who have intraepidermal nerve fibers and no pain, from people who have a slight decrease and they have pain, to people who have a disappearance of pain because of complete clearance of the nerve fibers. Do not do things that potentially may hurt.

Decompression of multiple nerves in diabetes are being advocated by some surgeons for the treatment of common or garden-variety DPN. There is a need to grade the severity of entrapment, use measures of sensory tests that have been validated, the presence of a Tinel’s sign, the presence of motor features and atrophy before considering surgery. A conduction block should be present, and there should not be severe neuropathy to preclude distinction from entrapment. Control and sham operations should be compared in order to determine outcomes from these procedures. Sham operations should be compared with unentrapment operations for symptoms if they are the only measure of the change, and objective standardized measures should be employed to show improvements in nerve function.

There may be a place for surgery in the treatment of common or garden-variety diabetic neuropathy. It is just that I have not been able to find it.

Aaron I. Vinik, MD, PhD, is director of the Strelitz Diabetes Research Institute and professor of medicine/pathology/neurobiology, Eastern Virginia Medical School, Norfolk, Va. He is associate medical editor of Diabetic Micorvascular Complications Today. He may be reached at vinikai@evms.edu or 757-446-5912.

For a downloadable pdf of this article, including Tables and Figures, click here.