Peripheral nerve fibers can be categorized according to their size and related functions. While large nerve fibers comprise myelinated Aα, and Aα/β fibers that control motor and sensory functions, the small nerve fibers comprise the unmyelinated C fibers and the thinly-myelinated Aδ fibers that control temperature perception, pain, and autonomic functions. In patients with small fiber neuropathy (SFN), both the Aδ and C fibers are damaged. Due to the diverse functions of small fibers, the symptoms of SFN are varied, with the most common clinical presentation including neuropathic pain (NeP), burning in the extremities, and autonomic dysfunction [1]. Increasingly more is being discovered on the mechanism of NeP in SFN. Transient receptor potential vanilloid subtype 1 (TRPV1) was first described to be involved in painful neuropathies in 2006 [2] and was later localized to the keratinocytes in patients with SFN [3]. Neurotoxins depleting TRPV1 induces the upregulation of multiple biochemical compounds and receptors, thus leading to a complicated cascade of cellular signaling changes [4].

Small fibre neuropathy (SFN) is a relatively common, but largely understudied neurological syndrome which has affected the lives of many globally. Although the estimated incidence of SFN is about 52.95 per 100,000 population [5], this is likely a gross underestimation due to diagnostic challenges. Studies even propose that SFN underlies 40% of fibromyalgia syndrome [6, 7] and hence, the actual number of sufferers could be close to 100 million worldwide [8].

While autoimmune diseases, diabetes, amyloid, sarcoidosis, toxic substances, and vitamin B12 (vit B12) deficiency are common secondary causes of SFN, no cause is found in around 50% of the cases [9, 10]. iSFN treatment often targets symptomatic relief using non-opioid and opioid analgesics, antidepressants, and antiepileptic medications [11]. Not only do these methods fail to address the underlying causes of iSFN, long-term use may also create drug dependence, specific drug-dependent adverse effects, and opioid abuse. Thus, the identification of biomarkers is recommended to screen for occult causes of initially iSFN (iiSFN) cases in patients without known risks. Identifying underlying causes in iiSFN patients can allow early treatment, thus greatly improving patients’ quality of life, particularly in young patients. As peripheral axons grow throughout life, early intervention in SFN patients may spur axonal regeneration which can improve or even cure patients’ symptoms [12]. A study of patients with mixed polyneuropathies obtained evidence showing that screening led to potentially disease-modifying management changes in 25% of them [12]. Hence, screening in patients with iiSFN for its most common occult causes is recommended [10, 13].

The prevalence and the debilitating yet non-specific effects of SFN underline the need for biomarker identification to stratify patients and develop cause-centered therapies. According to the Biomarkers, EndpointS and other Tools (BEST) glossary, a biomarker is a defining characteristic that is measured as an indicator of normal biological processes, pathogenic processes, or responses to an exposure or intervention, including therapeutic interventions. Biomarkers can be further divided into their functions: susceptibility/risk, diagnostic, monitoring, prognostic, predictive, pharmacodynamic/response, and safety [14]. In this way, SFN biomarkers can identify the patient population more accurately and may be used as diagnostic tests to tailor treatment plans according to different causes of SFN.

However, the “objectivity” of SFN biomarkers is affected by the local context and age of patients. Due to the complex nature of the disease, the frequencies of different causes of SFN may vary depending on geographical location, ethnicity, genetics, and environment. A study conducted on 84 patients in the Czech Republic showed diabetes mellitus (DM) as the most frequent and powerful risk factor for SFN in that region (odds ratio = 3.6, P = 0.009) [15]. In contrast, a study from the United States found that the most common etiologies of SFN were idiopathic (73%), presumed hereditary (18%), and DM (10%) [16]. The largest cohort study to date of 921 patients with pure SFN was conducted in the Netherlands, and it found that associated conditions in order of prevalence were autoimmune diseases, sodium channel (SCN) gene mutations, DM, glucose intolerance, and vit B12 deficiencies [10]. In addition, SFN commonly affects otherwise healthy children and young adults, in whom its association is often linked to inflammation or dysimmunity [17]. Hence, this review highlights the characteristics of various biomarkers and their functions to aid clinicians in discerning the types they should employ to diagnose and confirm SFN accordingly.

SFN is a disabling condition that significantly decreases a patient’s quality of life both physically and psychologically [11]. Due to the heterogeneity of the disease, diagnosis, search for etiology, treatment choices and prognostication are challenging. Therefore, biomarkers would significantly aid in these management aspects and the treatment of the root causes of SFN.

A literature review was performed with the search terms “small fiber neuropathy” and “biomarkers” or “diagnosis” or “diagnostic tools” or “diagnostic tests”, limited to human studies on PubMed/MEDLINE, Web of Science, and Embase from 2010 to the present. This year was chosen as the cutoff as guidelines from the European Federation of Neurological Societies/Peripheral Nerve Society were published on the use of skin biopsy and normative values of intra-epidermal nerve fiber density (IENFD) for bright field microscopy in the diagnosis of SFN [18]. This is to ensure stringent criteria of biopsy-proven SFN were used. The aim was to comprehensively summarize the important biomarkers to guide clinicians in the management of SFN in patients from various age groups. Ultimately, we hope that the information in this review may help reduce time, patient distress, and costs during the screening process and determine causal treatments of SFN to improve patients’ quality of life.

While SFN is an obviously recognizable diagnosis primarily diagnosed by history and clinical examination in patients with known risk factors such as diabetes, objective test confirmation is recommended for patients without known risks. The non-specific symptoms of SFN further emphasize the need for selecting biomarkers to exclude the occurrence of other neuropathies. Thus, diagnostic confirmation may be provided by functional neurophysiological tests and histological skin biopsies. However, it is worth noting that although these tests may help in diagnosing neuropathy, they may have limited value in assessing NeP in SFN [57, 58].

One functional neurophysiological test for the cardiovascular adrenergic function is the Valsalva maneuver (VM) test. The VM test evaluates baroreceptor sensitivity, as baroreceptors play a pivotal role in blood pressure regulation [56] and blood pressure recovery time (BPRT) is an often-used index of adrenergic baroreflex function. The patient with suspected SFN is rested and recumbent and asked to maintain a column of mercury at 40 mmHg by forcefully breathing out of their mouth with a tightly closed nose and straining for 15 s. Repeated tests are conducted with the patient tilted at a 20°–40° angle until a significant reduction in blood pressure is recorded [59]. In addition, the tilt table test measures changes in blood volume due to gravity redistribution in the transition to an upright angle between 60° and 80°. When standing from a supine position, excessive pooling of blood in the lower limbs is similarly prevented by baroreflex that is controlled by adrenergic outflow [60]. Positive tilt table test results may be present in patients with SFN due to abnormal changes in blood pressure, heart rhythm, or heart rate.

Amongst tests for the sudomotor function is the stimulated skin-wrinkling (SSW) test, which is a non-invasive indicator of limb sympathetic function [61]. SSW involves immersing the patient’s hands either in water or a EMLA, such as EMLA. The degree of wrinkling depends on vasoconstriction controlled by sympathetic nerve fibers in the glabrous hand and, to a lesser extent, foot skin [62]. EMLA-stimulated wrinkling showed 81.6% sensitivity and 74.7% specificity, whereas water wrinkling showed 71.4% sensitivity and 73% specificity in detecting abnormal IENFD in iSFN patients [63]. Studies have shown that compared to water wrinkling, EMLA produces a more linear response that persists for a longer time to allow for grading [62, 63]. As SSW is simple, cost-effective, and does not require specialized equipment, it is a practical clinical bedside test that has been implemented for routine use [64].

Other tests include quantitative sudomotor axon reflex testing (QSART) which measures the autonomic nerve fibers that control sweating to provide a quantitative assessment of postganglionic sudomotor function [65]. In the diagnosis of SFN, QSART has a sensitivity of 80% [66]. However, the test-retest reliability of QSART can be moderate with a large standard error of measurement [67], suggesting that the test is not ideally suited for longitudinal and interventional studies. Furthermore, QSART norms are only established for individuals who are 10 years or older [68] which excludes many suspected pediatric SFN patients. Hence, QSART should not be the only investigation modality used to confirm SFN.

In addition, QST assesses the psychophysical thresholds for cold and warm sensations, reflecting the function of the A–δ fibers. It is noninvasive, easily performed, and provides results that can be analyzed immediately to detect small fiber dysfunction [69]. A study found that QST is significantly more reproducible than QSART [70], so the two tests are often used in conjunction to improve the accuracy of results. Nonetheless, although QST has reasonable reproducibility, it has high inter-operator, inter-patient, and inter-test variability [71]. QST cannot differentiate between central and peripheral causes of temperature perception which means that any miscommunication between the examiner and patient would affect the results [72]. To improve the administration of QST in clinical practice and research, the Neuropathic Pain Special Interest Group of the International Association for the Study of Pain (NeuPSIG) recommended using predefined standardized stimuli and instructions, validated algorithms of testing, and established reference values corrected for the anatomical site, age, and gender [73]. When appropriate standards are applied, QST can provide valuable information about the functional status of the somatosensory system. Due to the disadvantages of QSART and QST, they are not recommended as stand-alone diagnostic confirmatory tests [74]. As such, the sensitivity of SFN diagnosis can be improved by combining the complementary assessments of somatic (QST) and peripheral autonomic small fiber function (QSART) [65]. The various autonomic and small fiber function tests that are clinically available for diagnostic confirmation of SFN following clinical examination are shown in Table 3.

There are other measures to assess sweat glands that are not validated for routine clinical use as of yet, such as the dynamic sweat test that assesses sweat gland density, distribution, and stimulated sweat production [75, 76]. To improve neurophysiological assessment by overcoming the inaccessibility of small nerve fibers with conventional nerve conduction tests, nociceptive laser-evoked potentials (LEPs) or contact heat-evoked potentials (CHEPs) can be recorded [77]. A recent study of Chinese adults found that CHEPs had a high reproducibility in diagnosing SFN [77]. Amplitudes of Aδ evoked potentials and leg skin flare responses from CHEPs were reduced in patients with sensory neuropathy compared to healthy controls (HCs) [78], and diagnostic accuracy was 74% compared to IENFD [79]. LEPs strongly associated with pinprick sensory disturbances at a sensitivity of 78% and specificity of 81% for the diagnosis of diabetic SFN [80].

Amongst multiple diagnostic tests available, skin biopsy has the highest accuracy to date in the diagnostic confirmation of SFN, in particular, a skin biopsy showing decreased IENFD [11]. In the past, objective confirmation of SFN required a surgical biopsy of a sensory nerve which is invasive, expensive, and hence only rarely performed [12]. Today, a skin biopsy is valued because it is minimally invasive and specifically assesses small fibers, unlike standard nerve conduction studies and electromyographic tests which are typically used to diagnose large fiber neuropathy [17]. Despite its advantages, skin biopsies are mainly limited by test insensitivity due to poor non-representative or published norms, especially in pediatric and female patients, sampling error, and symptoms typically appearing before axons degenerate [17].

Overall, skin biopsies have a sensitivity of 81–88% in diagnosing SFN [18]. Skin biopsy samples are first stained for protein growth product 9.5 (PGP 9.5, a pan-axonal antigen) which reveals small fiber axons with light or fluorescence microscopy [18]. Thereafter, these samples are sent to laboratories where they are processed and measured for IENFD relative to biopsies from healthy, demographically matched volunteers [18, 81, 82]. According to the European Federation of Neurological Societies, IENFD from a distal leg skin biopsy is reliable and efficient to confirm SFN, and laboratories should provide their own normative values, stratified for age, gender, and intra- and inter-observer reliability [81]. SFN diagnosis is confirmed in patients with a measured epidermal nerve density below the fifth percentile of the anticipated normal distribution. If false negative results occur, second biopsies (e.g., from the thigh for non-length-dependent neuropathic conditions or the foot [18, 83–86]) may add sensitivity but cost as well. There are also insufficient norms to recommend routine use.

Several studies have further improved the accuracy of skin biopsies for SFN diagnosis. A worldwide normative data set for IENFD of the distal leg by bright-field microscopy and indirect immunofluorescence corrected for age and gender has been established [82, 86]. Compared to conventional bright-field microscopy, immunofluorescence visualizes twice as many fibers and requires its own set of norms [81, 86]. Ongoing research into other biopsy sites and measures includes the density of sweat glands or arrector pili innervation, but more validation is needed before they can be implemented clinically [17].

In vivo corneal confocal microscopy visualizes the exclusively C-fiber innervation of the cornea is non-invasive and is effective for longitudinal tracking, but it is not yet a routine method [87]. More recently, SFN was identified using peripheral nerve ultrasound which shows an enlarged cross-sectional area of the sural nerve, although more research is needed before it can be implemented as a diagnostic method [88, 89].

If clinical examination, functional neurophysiological tests, and skin biopsies confirm SFN without a known secondary cause, these patients are labeled as “initially idiopathic” (iiSFN) and the next step would be to screen for occult causes [12]. Firstly, iiSFN patients are screened using medical history and prior test results. Secondly, hematological tests, lumbar punctures, or imaging are performed, identifying potentially treatable causes in 30% to 50% of iiSFN patients [10, 90, 91]. Thirdly, genetic sequencing is recommended for patients with a high probability of genetic causes. Therefore, the diagnostic performance and cost-effectiveness of specific tests must be defined for clinicians to decide on how to screen iiSFN patients for causes in an accurate yet optimized way.

Antibodies have long been associated with large fiber neuropathies and play a role in classifying the disease phenotype [100]. Although the pathophysiological mechanisms of SFN remain largely unknown, there is limited preliminary evidence showing that patients with SFN respond to immunotherapies and this correlates with the presence of auto-antibodies in SFN patients [101]. As SFN is a diagnostically and therapeutically challenging disease, further understanding of its pathophysiology is required to confirm its association with antibodies and to determine targeted treatments for SFN patients [102]. SFN may involve multiple immunological processes. Despite these SFN-associated autoantibodies, further validation and testing are required to determine pathogenicity for most of them.

Anti-sulfatide [immunoglobulin M (IgM)] antibodies have long been implicated in the sensory component in small and mixed fiber neuropathies. They are associated with acute onset in all SFN patients tested [103]. Steroid therapy was administered in the form of oral prednisone therapy to patients with acute SFN and there was a significant improvement in their clinical symptoms 1–2 weeks later. Of the four patients, three remained free of symptoms, and one experienced a recurrence of neuropathy after prednisone therapy ceased. However, these studies involve retrospective data analysis of clinical tests, and the level of evidence is low and merely associative [104].

Another study consisting of 155 patients with biopsy-confirmed iSFN was blindly tested for serum anti-tri-sulfated heparin disaccharide (TS-HDS) (IgM) antibodies. Of all the SFN patients tested, 48% had serum antibodies, and 37% specifically with anti-TS-HDS (IgM) antibodies. Compared to controls, anti-TS-HDS (IgM) antibodies were more frequent in SFN patients (P = 0.0012). Ninety-two percent of patients with acute-onset SFN had anti-TS-HDS (IgM) antibodies, which were more common in females and associated with NLD-SFN pathology. Hence, the presence of auto-antibodies in otherwise iiSFN patients’ sera, suggests that they may be used as markers for certain SFN subsets [104, 105].

There is also a suspected association between fibroblast growth factor receptor 3 (FGFR3) antibodies in the diagnosis of SFN, but its specificity remains unclear. In one study, 15% of iiSFN patients had serum anti-FGFR3 (IgG) antibodies present and in another, 17% had elevated serum anti-FGFR3 (IgG) antibodies [104, 105]. Although both TS-HDS/FGFR3 antibodies are found to be present in many SFN patients, more studies are needed to validate the role of these antibodies in the pathogenesis and clinical utility of immune-mediated dysautonomia.

Other novel autoantibodies for NeP include anti-plexin D1 antibodies. Plexin D1-IgG was originally detected in patients with NeP with underlying neuroinflammatory diseases, including atopic myelitis, multiple sclerosis, and neuromyelitis optica spectrum disorders [106]. Compared to control subjects, plexin D1-IgG antibodies were more frequent in patients with NeP (10% vs. 0%; P < 0.05). Furthermore, the application of anti-plexin D1 antibody-positive patient sera to cultured dorsal root ganglion (DRG) neurons increased membrane permeability and caused cellular swelling. This observation explains why patients with NeP and positive anti-plexin D1 antibodies commonly developed burning pain and current perception threshold abnormalities for C-fibers. Another study found that 75% of patients with iSFN present with a chronic persistent prickling or burning pain in a length-dependent manner [107]. There was a significant positive correlation between antiplexin D1-IgG levels with disease duration. This suggests that either a longer time is needed for SFN to fully develop in plexin D1-IgG, or that plexin D1-IgG may emerge due to long-term small fiber neuron damage because patients with secondary SFN caused by DM also carried plexin D1-IgG [108]. Although anti-plexin D1 (IgG) antibodies were confirmed to be pathogenic in SFN via mouse models, this study did not histologically confirm definite SFN in the test subjects [108].

One study confirmed an association between antibodies against SCNs, anti-Na_v (IgG) antibodies, and SFN in cell-based assays [109]. The patients assessed had an SFN variant of GBS and presented with acute onset burning sensations in a length-dependent distribution. These patients were treated with intravenous Ig and it was found that IgG antibodies in their acute phase sera bound strongly to murine small nerve fibers, but this binding ceased during the convalescent phase. Additionally, the study observed that serum IgG transfer to a murine nociceptive model caused transient alteration in thermal pain responses. Hence, this suggests that anti-Na_v (IgG) antibodies may induce an acute transient immune response directed at small nerve fibers [109].

More novel autoantibodies were recently reported by one study, in which researchers used a novel antibody technology [101] to identify putative autoantibodies in SFN using sera from SFN patients and age, gender-matched HCs that were screened against > 1,600 immune-related antigens on the patented array [110]. The results from this study elucidated that novel autoantibodies interferon-induced GTP-binding protein Mx1 (MX1), drebrin-like protein (DBNL), and cytokeratin 8 (KRT8) are found in iSFN, while MX1 in particular may enable diagnostic subtyping of iSFN patients as it was found to be higher in iSFN compared to secondary SFN. Hence, this study adds strength to the concept that an immune-mediated process may be involved in a subset of SFN. It provides a stepping stone for more studies to determine the pathophysiology of SFN for clinical subtyping and the development of targeted treatments for iSFN patients, who are currently treated symptomatically rather than causally.

A summary of the various auto-antibodies relating to their clinical significance and the level of evidence establishing their association with SFN is shown in Table 4.

The selection of appropriate biomarkers would improve patient stratification and management of SFN, whether secondary or idiopathic. As SFN patients often present with non-specific symptoms, clearer guidelines for SFN diagnosis would help clinicians to identify symptoms and implement tests that characterize SFN. However, awareness of the advantages and limitations of a given biomarker is important to minimize false positives and false negatives.

As outlined in this review, the diagnostic approach to SFN should be done in a stepwise manner, with consideration for each patient’s gender, age, ethnicity, and locality. Firstly, preliminary diagnosis can be made from clinical examination and history-taking, which usually identify patients with secondary causes of SFN such as DM. Secondly, diagnostic confirmation may be provided by functional neurophysiological and histological tests. Objective test confirmation is especially important for patients without known risk factors, because although they may present clinically with SFN symptoms, NeP may be due to other causes distinct from SFN. Thirdly, iiSFN patients without a known secondary cause are screened using blood tests and genetic sequencing to determine underlying etiology. Depending on the specificity of the patient’s symptoms and history, autoimmune tests can be administered for clinical subtyping if dysimmunity is suspected as an occult cause of iSFN.

By determining the etiology of iSFN, biomarkers can enable the development of better-targeted treatments that address specific causes of SFN. Significant breakthroughs have been made in developing and improving biomarkers of SFN, yet there is still much to uncover. We hope that this review encourages further studies to dispel the mystery surrounding SFN, and to enhance the diagnosis, treatment, and monitoring of this disease that has affected the lives of so many.