Utility of Skin Biopsy in Neuromuscular Disease

Recently, I asked some neurology residents if they would use intraepidermal nerve fiber density (IENFD) testing in future practice, and they said “Of course, how else do you diagnosis small fiber neuropathy (SFN)?” This got me thinking 2 things: whether I was very old, because I could easily recall diagnosing SFN without IENFD testing; and second, what was the appropriate role of IENFD in neurologic practice? This article reviews the utility of skin biopsy IENFD in diagnosis and management of neuromuscular disease.

The ability to visualize and quantify intraepidermal small nerve fibers from relatively noninvasive skin biopsy was a great step forward for research and clinical care for peripheral nerve disease.¹ Using an immunostain against protein gene product (PGP) 9.5, a nonspecific axonal marker, the small nerve fibers in the epidermis, consisting of unmyelinated C and thinly myelinated Aδ fibers, can be visualized and counted (Figure).

Before IENFD was available, SFN was diagnosed clinically or with autonomic testing.² Autonomic testing, however, is not readily available to most and may be normal in significant subsets of people with SFN. Clinical diagnosis may be sufficient in those with length-dependent numbness or dysesthesias and no findings of large fiber involvement on neurologic exam (ie, normal vibratory sense, proprioception, and deep tendon reflexes) and normal nerve conduction studies (NCS). Some argue IENFD testing adds little to clinical assessment. I believe this view too restrictive and hope to present a convincing case for the value of IENFD in specific circumstances.

Skin Biopsy Interpretation

Normal values for IENFD are available for several sites in upper or lower limb. The distal leg, 10 cm proximal to the lateral malleolus, is best studied for testing.^3,4 Samples should be matched for age and sex.⁴ By convention, results below the fifth percentile are considered abnormal. This is 2 standard deviations from the mean, so approximately 3% of individuals without SNF will have values in the abnormal range.

Results of IENFD are also considered abnormal if there is an increased number of axonal swellings (Figure B) or increased number of branched fibers. Although hematoxylin and eosin (H & E) staining is used to look for structural abnormalities, it seldomly contributes diagnostically useful information. Congo red staining for amyloid is also often routinely performed, and although it can identify amyloid in the skin for individuals with known familial amyloid polyneuropathy,⁵ the yield for pure SFN is low.

Skin Biopsy Clinical Use and Technique

Unlike conventional nerve biopsy (eg, sural nerve biopsy), skin biopsy results almost never reveal the cause of neuropathy. Results of IENFD testing typically only inform whether intraepidermal nerves are normal or not. In mixed fiber neuropathy (ie, small and large fibers both affected), which is the most common peripheral neuropathy, abnormal IENFD findings do not necessarily indicate pure SFN. If mixed fiber neuropathy is found clinically (eg, abnormal vibration or joint poisition sense, abnormal deep tendon reflexes, and abnormal NCS), there is generally no role for IENFD testing.

Skin biopsy is most useful when specimens are taken from more than 1 area of a person’s body to determine whether an SFN is nonlength-dependent or length dependent. In a length-dependent SFN, IENFD will be abnormal only in biopsies from distal sites or will be more abnormal in distal sites compared to proximal sites. Conversely, in nonlength-dependent SFN, IENFD is more affected in proximal than distal sites or may be multifocal. Skin biopsies show nonlength-dependent SFN in approximately one-third of cases.^6,7 Although this pattern is not diagnostic, patients with nonlength-dependent SFN are more likely to be younger women with immune-mediated disease processes.⁷ Individuals with SFN are also less likely to respond to treatment for neuropathic pain.⁷

Skin biopsy samples are obtained with a 3-mm circular punch tool. The procedure is easy to learn and several sites can be biopsied in a relatively short period of time. After cleansing target sites with alcohol and, optionally, adjunctive betadine, raise a small skin bleb using an insulin syringe with 1% lidocaine. The skin will quickly numb, and the punch can then be used to remove a plug of epidermis and subdermal fat. Cover the wound with simple adhesive bandage; sutures are unnecessary. Complications are infrequent and may include continued bleeding. Pressure with a small piece of folded gauze under the bandage to provide pressure will usually address this. Rarely, the biopsy site may become infected and require antibiotic treatment. People undergoing biopsy should be informed in advance that small scars may occur.

Biopsies are typically taken from sites for which normative data are available. A minimum of 2 sites, including distal and proximal areas should be sampled, and some advocate 3 sites. Biopsying more than 3 sites is rarely necessary. Although there are normative data for the foot (over the extensor digitorum brevis muscle), which is the most distal site available, there are concerns about sampling this area because no studies regarding reliability are available, and prior foot trauma could artifactually affect IENFD. Yield of biopsy from the distal leg is high, even in those with very distal symptoms.⁷ I prefer the distal leg and seldom biopsy the foot.

Other proximal sites that can be biopsied include the distal thigh above the knee and the proximal thigh below the hip. Of the 2 sites, the proximal thigh is more thoroughly studied. Biopsies can be taken from the wrist and upper arm (over the deltoid). Although normative values have been determined, upper extremity sites have not been studied with detail compared with lower limb sites. In my experience, even in people with predominantely upper limb symptoms, biopsy results are more often abnormal in the leg vs the arm.

There is usually no need to obtain biopsies from both sides of the body. Reasons to obtain bilateral biopsies include cases of markedly asymmetrical symptoms or when symptoms involve a part of the body for which normative data are not available. The symptomatic and asymptomatic sides can be compared; however, there is no standard for what degree of side-to-side asymmetry is abnormal.

Clinical Value of Skin Biopsy

As with any test, the role of that test in a given individual will depend on the clinical suspicion of disease and sensitivity and specificity of the test. Sensitivity and specificity of IENFD testing are very high, both around 90%,⁷ which means approximately 10% of people with an SFN will have normal IENFD results—a normal study does not exclude the possibility of SFN, but it decreases the likelihood markedly.

Because specificity is not 100%, a small percentage of normal individuals will have abnormal IENFD. If testing is performed too indiscriminately, false positive results can be expected. This leads to the key question of who should undergo IENFD testing (Table).

If there are length-dependent symptoms of numbness and no findings of large fiber involvement on neurologic examination or NCS, some clinicians will argue IENFD testing is not necessary. This conclusion is reasonable since a normal study would not exclude SFN as a potential diagnosis. Other clinicians prefer to obtain testing even in the setting of a high prior probability of SFN. Having an objective test is useful because up to one-third of patients with SFN can have normal sensory examination findings.⁸

I believe it best to have a conversation with the patient that addresses how test results will affect disease management. Other determining factors include cost to the individual being tested and her or his tolerance for invasive testing. If repeat IENFD testing in the future is a consideration, it might make sense to obtain testing at the time.

The greatest value of IENFD is for individuals with clinical presentations that are less typical for SFN (eg, intermittent, multifocal focal sensory symptoms). Before routine availability of IENFD testing, it was not appreciated that SFN can present with this pattern, which is now recognized as accounting from approximately 10% of cases.⁹ Other potential candidates for IENFD testing include people with myalgias but no evidence of myopathy or someone with apparent fibromyalgia. In a study, abnormal skin biopsy results were seen in 60% of people presenting with myalgias and mildly elevated creatine kinase levels.¹⁰ In a series of 233 people with suspected SFN confirmed by IENFD testing, 15% had pain described as “aching” in conjunction with typical pain characteristics (eg, numbness, tingling, and burning).⁶

When IENFD results are abnormal in people without typical features of SFN, different clinicians draw different conclusions. Some consider that the abnormal findings indicate the clinical spectrum of SFN is broader than originally appreciated (Box). Others are more circumspect and argue that abnormal IENFD testing in people without typical SFN features could be caused by mechanisms other than SFN or simply represents false-positive results. This debate is especially pertinent to the literature involving IENFD and fibromyalgia.

In studies of fibromyalgia and IENFD, among unselected people who met criteria for fibromyalgia but did not have symptoms of SFN, 30% to 50% had abnormal IENFD findings.¹¹ In 2 studies, 3% of age- and gender-matched healthy individuals had abnormally decreased IENFD, as would be expected because normal values are 2 standard deviations below the mean.^12,13 More than 30% of people with fibromyalgia, however, have decreased IENFD results, suggesting that this is not simply false-positive results.

It is unclear what these data mean. There is much research showing that central abnormalities play a role in fibromyalgia.¹¹ Are the small fiber changes seen in fibromyalgia secondary to central processes or are the central changes something that happens secondary to SFN? These questions remain unanswered and warrant further investigation.

False-positive rates were negligible in the studies discussed above; however, this was in research settings with IENFD testing performed at expert labs. An important but unanswered question is what is the false-positive rate in clinical practice? Numerous studies have suggested the specificity of IENFD testing for SFN is around 90%, indicating a false-positive rate of 10%.⁷ Among experienced labs and readers, there was good reproducibility of IENFD measurements when the same sections were processed at several different labs and interpreted by different readers.¹⁴

Assuming the labs performing IENFD testing are employing optimal techniques for processing and quantitation,³ the main variable is the handling of the skin tissue during biopsy and transport to the lab. Like any procedure, the punch biopsy procedure has a learning curve.; however, this is not very steep. Among the most common errors made by individuals new to this procedure is not advancing the punch deep enough. At almost every site, there is no danger from advancing the metal portion of the punch all the way into the skin. Doing this ensures the epidermis is completely cut and that subdermal fat will be affixed to the specimen. Therefore, there is no need for forceps to tug on the specimen and the specimen can be grasped via the subdermal fat. If the epidermis is pulled or pinched, then artifacts can be introduced that could falsely decrease the IENFD measurements obtained.

Some question the accuracy of a lab’s results because almost all the cases they submit have abnormal IENFD. Of course, the percentage of cases coming back abnormal will depend on the cases being biopsied, so no 2 clinicians will necessarily have the same rates of abnormal studies. A study of 233 patients evaluated by neuromuscular specialists at 3 different centers found that approximately 70% of people evaluated for SFN by skin biopsy had abnormal results.⁶ This percentage did not significantly differ among individual centers or neurologists.

Assuming the lab is doing an appropriate job, if someone is having greater than 90% of their skin biopsy cases coming back abnormal, this suggests they are biopsying almost exclusively abnormal patients or they are submitting suboptimal specimens, introducing artifacts during the biopsy procedure. If your biopsy technique is appropriate, then you should consider whether you should be biopsying a larger selection of your patients. Because approximately 10% of people with SFN will have normal IENFD (because the sensitivity is only around 90%), if almost all of your biopsy results are coming back abnormal, a problem with the biopsy technique or the lab seems the most likely explanation.

It is important to appreciate that an abnormal skin biopsy result does not equate to a diagnosis of SFN. As with any diagnostic test, the results need to be analyzed in the context of the history, neurologic examination, and the results of other tests. Although IENFD testing is a useful diagnostic tool, its limitations must always be kept in mind.

Other Roles for Skin Biopsy

Aside from the diagnosis of SFN, IENFD testing can be helpful in other ways. My colleagues and I have used serial skin biopsies to demonstrate objective improvement in patients with suspected immune-mediated SFN treated with immunotherapy.^15,16 Repeat skin biopsies taken 6 months later from sites a few millimeters from the original biopsies showed significant increases in IENFD in most patients. This is important because other outcomes assessments are mainly subjective. Other authors have shown that IENFD can improve with treatment of immune-mediated neuropathies.¹⁷ In the natural history of idiopathic SFN, there is progressive decrease in IENFD; improvement is not seen.¹⁸ Certainly, further study of this role for IENFD testing is warranted.

Other Tests for Small Fiber Neuropathy

Other tests can provide objective diagnosis of SFN, including autonomic testing and quantitative sensory testing (QST). The most sensitive part of autonomic function is quantitative sudomotor axon reflex testing (QSART), which showed sensitivities of 59% to 80% .¹⁹ The limited sensitivity is not surprising because not everyone with SFN will have autonomic involvement. Some patients with symptoms of SFN but normal IENFD may have abnormal QSART.²⁰ The sensitivity of QST ranges from 60% to 85%.¹⁹

The greatest limitation of QSART and QST is that these tests are typically only available at specialized centers. Other modalities that have been studied to diagnose SFN, such as corneal confocal microscopy, laser-evoked potentials (LEPs), and contact heat-evoked potentials (CHEPs) are even less accessible.

Several labs that provide IENFD testing also offer testing of sweat gland nerve-fiber density (SGNFD). This is often used to improve the yield of skin biopsy testing. If IENFD test results are normal then decreased SGNFD is used as evidence for an SFN. Caution must be urged for this approach because there are no published data addressing this strategy. Indeed, currently offered SGNFD testing is based on a single study of 30 participants with diabetic neuropathy and 64 healthy participants.²¹

Conclusions

Readily available to all clinicians, IENFD is a valuable test. The skin biopsy procedure is well-tolerated, has few complications, can be learned easily, and can be performed in an office or clinic. As with any diagnostic test, care must be taken to avoid technical artifacts and appropriate individuals must be selected for testing. The literature demonstrates excellent sensitivity and specificity, supporting the role of IENFD testing in the diagnosis of SFN. It is hoped that further research will better define the role of IENFD testing in assessing fibromyalgia and other pain states as well as the role of serial testing.