Laboratory Evaluation of Peripheral Neuropathy

K.H. Vincent Lau, MD


Semin Neurol. 2019;39(5):531-541. 

In This Article

Chronic Peripheral Neuropathy

Chronic Distal Symmetric Predominantly Sensory Polyneuropathy

The most common type of peripheral neuropathy is chronic, distal, symmetric, predominantly sensory polyneuropathy (DSP),[1] presenting with numbness and tingling insidiously in a length-dependent (distal to proximal) fashion. Neurophysiological testing may not be required if the presentation is classic, but would reveal a generalized, length-dependent, primarily axonal polyneuropathy, with sensory abnormalities greater than motor. The broad differential diagnosis includes diabetes mellitus, vitamin deficiency (B12 or E) or toxicity (B6), hypothyroidism, paraproteinemia (benign monoclonal gammopathies, multiple myeloma, Waldenstrom's disease), connective tissue diseases (systemic lupus erythematosus, rheumatoid arthritis), infections (e.g., human immunodeficiency virus [HIV]), medication-induced (e.g., chemotherapy such as platinum-containing compounds, antibiotics such as metronidazole or dapsone), toxin-induced (e.g., alcohol),[12] and heavy metal toxicity or deficiency (e.g., copper deficiency).[13] From this list, it is clear that the initial step to narrowing the differential diagnosis is taking a careful history. An older study suggested that the combination of history, physical examination, and laboratory testing can identify the underlying etiology of DSP in at least 73% of cases,[14] which may be greater in this era of more sophisticated laboratory testing.

Classically, initial screening laboratory studies for DSP include fasting blood glucose, vitamin B12 level, paraprotein screen, thyroid function tests, full blood count, liver function tests, and erythrocyte sedimentation rate (ESR).[12] This panel continues to be advocated by the American Academy of Family Physicians.[15] The aforementioned Practice Parameter published by the AAN, AANEM, and AAPM&R[1] emphasized that among these tests, those with greatest yield is blood glucose—simply because diabetes mellitus is the most common cause of DSP—followed by serum vitamin B12 and serum protein immunofixation electrophoresis (SPEP).

In the clinical practice of testing for diabetes or prediabetes, hemoglobin A1c is ordered most frequently, followed by fasting glucose. In one study, glucose tolerance test (GTT) was ordered in only 1% of patients with peripheral neuropathy,[9] even though prediabetes, or impaired glucose tolerance, has been associated with DSP especially in patients in which pain is a prominent symptom.[16] As such, the recent Practice Parameter[1] recommended fasting blood glucose rather than hemoglobin A1c as most appropriate for screening, with the recommendation of GTT for patients with severe pain.

Vitamin B12 level is also checked frequently in patients with DSP, second only to hemoglobin A1c.[9] The serum vitamin B12 test measures the sum of haptocorrin-bound and transcobalamin-bound molecules, even though only the latter is taken into cells to meet metabolic demand.[17] Vitamin B12 is required for methionine synthase to remethylate homocysteine to methionine, so elevated homocysteine is a marker for vitamin B12 deficiency, with the caveat that folate should be tested simultaneously as homocysteine elevation may also reflect folate deficiency. Similarly, methylmalonyl-CoA mutase uses adenosylcobalamin to convert methylmalonyl-CoA to succinyl-CoA; in vitamin B12 deficiency, excess methylmalonyl-CoA is instead hydrolyzed to methylmalonic acid (MMA), so its elevation is also suggestive of vitamin B12 deficiency if renal function is normal. The role of both elevated homocysteine and elevated MMA to confirm vitamin B12 deficiency in the setting of borderline serum levels is well accepted, as the improved sensitivity has been shown to be clinically relevant.[18] A recent small series suggested a relationship between elevated homocysteine and sensory polyneuropathy even with normal vitamin B12, folate, and MMA.[19] As such, borderline vitamin B12 should be followed with MMA and homocysteine, interpreted alongside folate level and renal function.

SPEP is checked less frequently than vitamin B12 or hemoglobin A1c.[9] In a process called electrophoresis, SPEP separates out and measures serum proteins called immunoglobulins,[8] classified for their heavy chain subtypes (more commonly IgG, IgM, or IgA, and less commonly IgE or IgD) and light chain subtypes (kappa or lambda). The test evaluates whether one type of immunoglobulin is produced in excess due to abnormal clonal proliferation of B-lymphocytes or plasma cells in what is termed "monoclonal gammopathy," seen in the conditions of multiple myeloma, lymphoma, amyloidosis, Waldenstrom's macroglobulinemia, cryoglobulinemia, and POEMS (polyneuropathy, organomegaly, endocrinopathy, M-protein spike, and skin manifestations). However, abnormal immunoglobulins are most frequently seen in MGUS, which is estimated to be present in approximately 3.2% of all individuals older than 50 years.[8] For detecting monoclonal gammopathies, in general, SPEP is usually performed as an initial investigation, which identifies the heavy chain subtype. This is followed by immunofixation (IFE) to confirm the presence of abnormal immunoglobulin and to further characterize both heavy and light chain subtypes.[20] IFE has superior sensitivity (87%) for monoclonal gammopathies compared with SPEP (79%), although it is more expensive and labor intensive.[21] It also offers additional light chain analysis needed to characterize specific conditions such as light chain amyloidosis.[20] In practical terms, an abnormality noted on SPEP or IFE generally requires further investigation with referral to a hematology specialist for further workup, which may include bone marrow aspiration.

Other laboratory investigations should be pursued in the presence of associated symptoms and risk factors. These include connective tissue disorders, for example, rheumatoid arthritis and systemic lupus erythematosus[22] for patients with related complaints such as joint pain, evaluated with antinuclear antibody (ANA), extractable nuclear antigen (ENA), and rheumatoid factor (RF). Infectious causes including Lyme's disease, HIV, and hepatitis may be evaluated in patients with possible exposure history;[23] DSP is common among those with HIV (estimated 20–60%) and less so hepatitis, for which patients often have concurrent mixed cryoglobulinemia.[24] Vitamin levels may be checked—pyridoxine for patients with possible excessive supplementation and thiamine and vitamin E deficiencies for patients with malnutrition or absorption abnormalities. Heavy metals can also be checked for patients with exposure history, including copper and zinc. Sarcoidosis can be evaluated with serum angiotensin-converting enzyme (ACE) or cerebrospinal fluid (CSF), although specificity is limited and further supportive clinical evidence is often necessary.[25]

The yield of testing without a corresponding history is likely to be low, although, as previously discussed, no formal study has proven this. Based on a prospective study, if initial diagnostics do not yield an etiology, repeat laboratory testing is unlikely to elucidate a cause at a later time.[26] Nerve biopsy is considered rarely helpful in slowly progressive, chronic polyneuropathy.[27]

Chronic Primary Demyelinating Polyneuropathy

Demyelinating polyneuropathies generally present with atypical features that prompt neurophysiological testing, for example, proximal weakness in chronic inflammatory demyelinating polyneuropathy (CIDP) or severe sensory ataxia in anti–myelin-associated glycoprotein (MAG) peripheral neuropathy. The differential diagnosis includes CIDP, multifocal motor neuropathy (MMN), demyelinating variants of Charcot-Marie-Tooth (CMT) disease, anti-MAG peripheral neuropathy, POEMS, and GALOP (gait ataxia, autoantibodies IgM GD1a autoantibody against central myelin antigen "galopin," late age of onset greater than 70, polyneuropathy with symmetric and sensory greater than motor features).[28,29] Notably, there are etiologies that can be elicited on history rather than laboratory testing, for example, demyelinating polyneuropathy as an adverse effect of tumor necrosis factor-α antagonists.[30]

CIDP and MMN are easily differentiated by presentation, the former presenting with proximal, generally symmetric symptoms and the latter presenting with multifocal neuropathies. Neurophysiological findings often include conduction block. For CIDP, the supportive laboratory finding is elevated CSF protein (>0.45 g/L) without pleocytosis, seen in over 90% of patients.[31] Antibody screening is not routinely performed in part due to pathophysiological heterogeneity. In one study of 65 CIDP patients, 18.6% had anti-ganglioside antibodies with 6.2% against contact-1, 4.6% against neurofascin 155, 1.5% against contactin 1/contactin-associated protein-1 complex, and 1.6% against peripheral myelin protein.[32] Although less useful diagnostically, antibodies may have prognostic value. For example, neurofascin IgG4 antibodies are associated with poorer response to IVIG as well as clinical development of disabling tremor,[33] while antibodies to LM1 and LM1-containing complexes may be associated with ataxia.[34] Nerve biopsy in CIDP is rarely performed given the availability of other diagnostic criteria, although it may be helpful for atypical cases with equivocal neurophysiological findings without albuminocytologic dissociation.[35]

As for MMN, the supportive laboratory finding is the presence of anti-GM1 antibodies, with sensitivity estimated around 40 to 67% and specificity 85 to 91%. The addition of galactocerebroside to the anti-GM1 assay to detect anti-GM1–GalC complex was recently found to increase sensitivity to 65 to 81% with some compromise to specificity at 80 to 91%.[36,37] The addition of NS6S to anti-GM1 antibodies raised the combined sensitivity to 64% in one study, although specificity was limited.[38] Other antibodies including GM2 and galactocerebroside have also been found to have low specificity and positive predictive value for MMN.[39]

For CMT, although some patients may present with the "classic" presentation of pes cavus and "stork legs," many subtypes present atypically. It is difficult to quantify the yield of genetic testing for CMT because the "hit rate" depends on pretest suspicion. For example, the yield of testing for CMT1A (PMP22 duplication) was found to be high when selecting patients with an autosomal dominant family history and primary demyelinating polyneuropathy on neurophysiological studies (54–80%), but much lower when tested indiscriminately (25–59%).[1] As such, the recent aforementioned Practice Parameter[1] recommends a nuanced algorithm for testing. If there is a positive family history, the combination of inheritance pattern (autosomal dominant, autosomal recessive, or X-linked) and neurophysiological features (demyelinating or axonal) determines the first tier genes to be tested, followed by a second and then third tier. For example, in a patient with a positive family history suggestive of autosomal dominant inheritance and demyelinating features on neurophysiological studies, the algorithm suggests first testing for PMP22 duplication (70% yield), followed by MPZ mutation (5%), and PMP22 mutation (2.5%), followed by EGR2 mutations or LITAF mutation.

Since the publication of the Practice Parameter, a different algorithm was proposed based on a large-scale study at Wayne State University School of Medicine.[40] It was found that in 1,024 patients diagnosed with CMT, five subtypes accounted for 90% of diagnoses—CMT1A and HNPP (PMP22), CMT1B (MPZ), CMT1X (GJB1), and CMT2A (MFN2). A more recent cross-sectional analysis based on the Inherited Neuropathies Consortium natural history study analyzing 1,652 patients across 13 centers found a very similar result that these five subtypes accounted for 89.2% of enrolled patients.[41] In the former study, 67% of patients with a clinical CMT diagnosis had genetic confirmation, compared with 60.4% in the latter. These investigations support the recommendation that it may be more cost-effective to first study the four genes accounting for five subtypes, and pursue further testing only if this initial panel is negative. Alternatively, some genetic testing companies offer tiered testing, for example, allowing clinicians to send a subpanel of CMT testing for axonal or demyelinating subtypes, with the option to test the remainder of the panel if the first screen is negative.[42] Current practice in genetic testing depends largely on clinician experience and the availability of panels by testing companies.

In anti-MAG peripheral neuropathy, patients generally present primarily with sensory complaints such as distal paresthesias, sensory ataxia, and impaired proprioception, alongside varying degrees of distal muscle weakness.[43] The demyelinating features on NCS specifically demonstrate markedly prolonged latencies out of proportion to slowed conduction velocities. SPEP and IFE may identify abnormal IgM spikes. The definitive test for diagnosis is anti-MAG antibodies, with ELISA preferred over Western blotting due to easier technique without sacrificing sensitivity. In one study, anti-MAG antibodies were found in 72% of patients with demyelinating polyneuropathy and IgM monoclonal gammopathy.[44] In POEMS, distal sensory symptoms also present prior to motor symptoms, with approximately half of patients developing severe weakness. Laboratory testing often reveals elevated serum vascular endothelial growth factor (VEGF) several times above the normal limit.[45] In GALOP, disabling gait disturbances and sensory abnormalities develop in elderly patients.[29] Antibodies to sulfatide and galopin may be positive. Note that these conditions are difficult to diagnose, as motor involvement is variable, so patients may present similarly with cases of DSP. Clinicians must be careful to note atypical features on history and examination that may prompt neurophysiological testing, which may ultimately reveal demyelinating polyneuropathy.

In practice, in the presence of conduction block on neurophysiological testing, CSF studies and anti-GM-1 testing may be considered first line. Depending on the clinical presentation, testing for anti-MAG antibodies and VEGF should also be pursued, with sulfatide and galopin considered. With a positive family history, patients should undergo genetic testing for demyelinating forms of CMT.

Chronic Motor Predominant Axonal Polyneuropathy

Motor predominant neuropathies have a clinical phenotype characterized by muscle weakness (distal or proximal), muscle atrophy, and hyporeflexia out of proportion to sensory involvement. Early or severe motor involvement beyond that seen in DSP should prompt neurophysiological testing, which would demonstrate motor abnormalities with absent or only mild sensory abnormalities. The differential diagnoses for demyelinating polyneuropathies with variable motor involvement were discussed earlier. For axonal polyneuropathy with motor predominance, the narrow differential diagnoses include axonal CMT and distal hereditary motor neuropathy (dHMN). Genetic testing for CMT was discussed in the previous section. For dHMN, the most common genetic abnormalities are HSPB1 (dHMN type 1), HSPB8 (dHMN type 2), and BSCL2,[46,47] which, similar to CMT, may be found in genetic panels offered by testing companies.[42] For chronic presentations of motor predominant motor neuropathies, it is reasonable to start with one or both genetic testing panels, even if there is no corresponding family history. Acute or subacute causes of motor axonal polyneuropathies are discussed separately.

Chronic Pure Sensory Neuropathy/Neuronopathy

Pure sensory neuropathy, also known as sensory neuronopathy or ganglionopathy, presents with severe sensory complaints including sensory ataxia and early proprioception loss, greater than that expected for DSP. Neurophysiological testing reveals abnormalities only in sensory responses; a recent study suggested that sensory nerve action potential asymmetry of more than 50% in two pairs of nerves (median, ulnar, radial, sural, and superficial peroneal) has a positive predictive value of 94.3% for sensory neuronopathy.[48]

The differential diagnosis for sensory neuronopathy is relatively narrow, and as many are treatable, it may be reasonable to test for all the acquired causes. These include Sjogren's syndrome,[49] nonceliac gluten sensitivity and celiac disease,[50] anti-Hu and anti-CV2/contactin response mediator protein 5 paraneoplastic syndromes,[51] vitamin B6 excess,[52] and HIV.[53] An inherited cause is Friedrich's ataxia, typically associated with profound sensory complaints but with more accompanying features alongside neuropathy.[54]

Laboratory screening for Sjogren's syndrome includes testing for Sjogren's syndrome A or B (SS-A or SS-B) antibodies. If these are negative but the clinical suspicion remains high based on sicca symptoms and a painful, distal, sensory neuropathy, lip biopsy may be considered. In one study, anti-SS-A and anti-SS-B were detected in only 30% of patients who later had lip biopsies suggestive of Sjogren's syndrome, although this result is skewed, as those with positive antibodies would not proceed to have lip biopsies.[55] It has been estimated that 5 to 15% of patients with Sjogren's syndrome have some form of neuropathy.[56] Laboratory testing for gluten sensitivity includes tests for antigliadin IgG or IgA, endomysial and transglutaminase-2 antibodies. It was recently suggested that transglutaminase-6 antibodies may be helpful in identifying gluten sensitivity without definitive celiac disease.[50] Screening for paraneoplastic disease is generally ordered as part of a serum panel. Screening for HIV involves immunoassays that detect p24 antigen along with IgM and IgG antibodies to HIV-1 and -2. If the screen is positive, a follow-up assay differentiates between HIV-1 and HIV-2 viruses, and if these are negative, nucleic acid testing for HIV-1 is performed.[57] Given the relatively narrow differential and marked neurophysiologic features of this type of peripheral neuropathy, most of these investigations should be considered as screening.

Friedrich's ataxia is diagnosed with genetic testing for trinucleotide repeat,[58] suspected in children and adolescents with profound gait abnormalities and possible family history.

Small Fiber and Autonomic Neuropathy

Laboratory testing for small fiber neuropathy (SFN) is divided into studies that diagnose SFN and those that aim to elucidate the underlying etiology. SFN is more difficult to diagnose and recognize than the aforementioned types of peripheral neuropathy, as neurophysiological testing is unrevealing. The classic presentation of SFN is length-dependent loss of sensation to thermal or pain modalities, with or without prominent autonomic complaints such as impaired sweating, erectile dysfunction, gastroparesis, constipation, and urinary retention.[59] Due to low specificity of some of these symptoms, history taking may be assisted by questionnaires such as the Small Fiber Neuropathy and Symptoms Inventory Questionnaire[60] or Composite Autonomic Symptom Score-31,[61] and examination may be assisted by standardized instruments. Despite these, it remains helpful to have supportive laboratory testing to confirm the suspicion of SFN.

Skin biopsy allows for the quantification of somatic and autonomic small nerve fibers. SFN is diagnosed if intraepidermal nerve fiber (IENF) density falls below age- and sex-adjusted normative reference values.[62] The quantification of IENF is performed manually by a pathologist by counting, although there is recent interest in automating the method for increased accuracy.[63] Skin biopsy is generally performed on both a proximal and distal site to differentiate between SFN—expected to have greater differential between the two sites—and ganglionopathy—expected to have more similar values.[64] At one institution, skin biopsy performed for possible SFN was found to change diagnosis or management in 52% of patients.[64] Overall, the sensitivity is estimated to be 74 to 90% and specificity 64 to 90%.[65] The absolute "yield" of skin biopsy and similar tests is difficult to interpret as these are related to pretest probability, although skin biopsy is generally well accepted for the diagnosis of SFN.

In contrast, quantitative sensory testing (QST) is used less commonly in clinical practice, in part due to the time required to perform the procedure.[59] QST involves applying specific calibrated stimuli and recording subjects' response.[66] For suspected SFN, thermal and mechanical stimuli are applied in a graded fashion while patients report when sensations are felt. Clearly, there is a degree of subjectivity influencing results, along with patient motivation, attention, cognitive ability, and potential for malingering.[66,67] QST also has no localizing value as abnormalities may be caused by lesions anywhere in the sensory neuroaxis from peripheral nerves to the cortex.[67] The sensitivity of thermal QST for SFN has been reported to range from as low as 36% to as high as 100% against skin biopsy-confirmed diagnosis as the gold standard.[66]

The quantitative sudomotor axon reflex sweat test (QSART) evaluates autonomic nerve integrity by measuring postganglionic sympathetic cholinergic function, using sweating as a marker.[68] The test involves applying an electrical stimulation in a process called iontophoresis to introduce acetylcholine into the skin, with subsequent measurement of the volume of sweat produced. The reliability of the test has been questioned.[69] A test similar in principle to QSART is electrochemical skin conductance (ESC), which also evaluates sweat function. A recent systematic review found significant variability in ESC values between diseases, and longitudinal studies of disease demonstrated ESC changes that were not physiologically possible, questioning test validity.[70]

Various diagnostic criteria for SFN based on QST and QSART have been proposed, for example, that a diagnosis of SFN may be considered definite with abnormal neurological examination (with abnormal sensation to pinprick) and abnormalities on both QSART and QST; probable with abnormal neurological examination and either abnormal QSART or QST; and possible with one of abnormal neurological exam, QSART or QST.[71]

Heart rate variability (HRV) has also been proposed as an indirect way to assess autonomic nervous system tone, which can be measured at the bedside or using more sophisticated analyses including time, frequency domain, and nonlinear analysis. It has limited specificity, however, as abnormalities may also reflect cardiovascular disease.[72] Overall, the combination of clinical history, neurological examination, and skin biopsy may be considered first line in the diagnosis of SFN, followed by QST and QSART.

After the diagnosis of SFN is made, the evaluation of an underlying etiology is similar to that for DSP, with a very broad differential diagnosis. The most common etiologies overlap with those for DSP, specifically diabetes mellitus, vitamin abnormalities, and monoclonal gammopathy of unknown significance. An important comprehensive review[58] suggests the following tiered testing: an initial screen may include tests for diabetes mellitus or prediabetes with hemoglobin A1c, fasting glucose and 2-hour GTT, vitamin abnormalities with vitamin B12 level, and MMA and paraproteinemia with SPEP and IFE. Other initial tests may include triglyceride levels, thyroid-stimulating hormone, and vitamin B6, B1, and E. A second-line evaluation may include tests for SS-A and SS-B, Lyme's disease (ELISA or Western blot), sarcoidosis (serum ACE), heavy metals, and HIV, while a third-line evaluation may include amyloidosis (genetic testing or fat aspirate), Fabry's disease (α-galactosidase A assay), hepatitis C (virus antibody screening, ribonucleic acid confirmation if screen positive), leprosy (biopsy of the lesion), paraneoplastic disease (anti-Hu, anti-CV2/collapsing response mediator protein 5, ANNA-3, anti–leucine-rich glioma inactive protein 1), acute intermittent porphyria (porphobilinogen in urine and erythrocytes), celiac disease (antigliadin antibodies), and other hereditary neuropathies (genetic testing).[59]

Although autonomic dysfunction is described in the context of SFN here due to considerable overlap, testing for primary autoimmune autonomic neuropathy via α3 gAChR[73] or hereditary sensory autonomic neuropathy[74] may be appropriate in less common scenarios of isolated autonomic dysfunction without other sequelae of SFN.