*Review* **Differential Diagnosis of Acquired and Hereditary Neuropathies in Children and Adolescents—Consensus-Based Practice Guidelines**

**Rudolf Korinthenberg 1, \* , Regina Trollmann 2 , Barbara Plecko 3 , Georg M. Stettner 4 , Markus Blankenburg 5 , Joachim Weis 6 , Benedikt Schoser 7 , Wolfgang Müller-Felber 8 , Nina Lochbuehler 9 , Gabriele Hahn <sup>10</sup> and Sabine Rudnik-Schöneborn 11**


**Abstract:** Disorders of the peripheral nerves can be caused by a broad spectrum of acquired or hereditary aetiologies. The objective of these practice guidelines is to provide the reader with information about the differential diagnostic workup for a target-oriented diagnosis. Following an initiative of the German-speaking Society of Neuropaediatrics, delegates from 10 German societies dedicated to neuroscience worked in close co-operation to write this guideline. Applying the Delphi methodology, the authors carried out a formal consensus process to develop practice recommendations. These covered the important diagnostic steps both for acquired neuropathies (traumatic, infectious, inflammatory) and the spectrum of hereditary Charcot–Marie–Tooth (CMT) diseases. Some of our most important recommendations are that: (i) The indication for further diagnostics must be based on the patient's history and clinical findings; (ii) Potential toxic neuropathy also has to be considered; (iii) For focal and regional neuropathies of unknown aetiology, nerve sonography and MRI should be performed; and (iv) For demyelinated hereditary neuropathy, genetic diagnostics should first address PMP22 gene deletion: once that has been excluded, massive parallel sequencing including an analysis of relevant CMT-genes should be performed. This article contains a short version of the guidelines. The full-length text (in German) can be found at the Website of the "Arbeitsgemeinschaft der Wissenschftlichen Medizinischen Fachgesellschaften e.V. (AWMF), Germany.

**Keywords:** neuropathy; children; adolescents; Charcot–Marie–Tooth disease; traumatic neuropathy; inflammatory neuropathy; metabolic neuropathy

**Citation:** Korinthenberg, R.; Trollmann, R.; Plecko, B.; Stettner, G.M.; Blankenburg, M.; Weis, J.; Schoser, B.; Müller-Felber, W.; Lochbuehler, N.; Hahn, G.; et al. Differential Diagnosis of Acquired and Hereditary Neuropathies in Children and Adolescents— Consensus-Based Practice Guidelines. *Children* **2021**, *8*, 687. https:// doi.org/10.3390/children8080687

Academic Editor: Mark S. Scher

Received: 28 June 2021 Accepted: 5 August 2021 Published: 9 August 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

Peripheral neuropathies are among the more frequent diseases confronting neurologists in their daily practice. The spectrum of aetiologies, clinical presentations, and disease courses is very broad, and differs considerably depending on the patient's age. Differential diagnostics for children and adolescents can be especially challenging because the more frequent neuropathies affecting adults (the diabetic, alcoholic, and vascular forms) very seldom (if at all) affect children, whereas rare hereditary and metabolic syndromes reveal a vast aetiological spectrum. This situation is made even more difficult because electrophysiological examination methods are often distressing for children, and only a few paediatricians are trained in electrophysiological methods.

These guidelines have been drafted for a wide variety of paediatricians and specialists (neuropaediatricians, neurologists and clinical neurophysiologists, genetic counsellors and consulting services, paediatric neuroradiologists, neuropathologists, and paediatric metabolic specialists) to provide orientation regarding diseases of the peripheral nerves affecting children and adolescents. The aim of these guidelines is to describe the state-ofthe-art differential diagnoses and consensus-based target- and cost-oriented diagnostics.

#### **2. Materials and Methods**

The methods applied for these guidelines follow the regulations of the Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften (AWMF; Version 1.1, 27.02.2013) [1]. The basis for the present version of guidelines was a non-systematic review of the recent literature by the coordinating author and by the co-operating specialists for their respective fields of expertise. All these specialists also have vast practical experience in their fields, so their clinical perspectives in addition to economic aspects have enriched these guidelines. As the Delphi technique requires, the diagnostic recommendations were considered and consented to by the whole working group in a multi-step approach. In the first step, the coordinating author collated proposals from the group members, put them into written form, and returned them to the group. Each member could then agree to a recommendation or express another opinion to be discussed. Next, the coordinating author collected all the members' feedback and presented their discussion points in an anonymous form. Those were then sent back to the group members, who reconsidered each recommendation. This process was repeated until a solid consensus on each recommendation was achieved.

The power of a recommendation was classified in 3 levels, each with its own designation [1]:


The strength of consensus was classified as [1]:


We achieved a "strong consensus" or "consensus" on all recommendations after the third round of consideration. Additionally, the text and comments of the guidelines were informally optimized by the guideline group through several voting sessions. Finally, the boards of all the participating scientific societies gave their final approval of the guidelines. The full-length text of the guidelines (in German) can be found at https://www.awmf.org/ leitlinien/detail/ll/022-027.html (accessed on 6 August 2021).

#### **3. Definition and Classification of Neuropathies in Children and Adolescents**

Neuropathies are diseases of the peripheral and cranial nerves, whose anatomical and function-bearing structures consist of axons and myelin sheaths, endo-, peri-, and

epineural connective tissue, and the vasa nervorum. Hereditary, traumatic, malignant, inflammatory, vascular, and metabolic disorders can cause damage to these structures. They can affect many nerves (polyneuropathy), or individual nerves (mononeuropathy, mononeuritis multiplex). Table 1 contains an overview.

**Table 1.** Survey of neuropathies in children and adolescents.


DSMA: distal spinal muscular atrophy; HNPP: hereditary neuropathy with pressure palsies; HNA: hereditary neuralgic amyotrophy; AD: autosomal dominant; AR: autosomal recessive; XD: X-chromosomal-dominant; XR: X-chromosomal recessive. Comprehensive names for groups of diseases are printed in bold.

> **Nerve injuries** occur due to sharp or dull mechanical effects or tearing. A functional conduction disorder without a transected axon is termed "neurapraxia", which heals relatively quickly. A transected axon (but with intact adjacent structures) is termed "axonotmesis". Here, recovery is usually achieved by sprouting from the proximal axon end. In cases of "neurotmesis", the entire nerve's continuity is broken, frequently resulting in a scar neuroma; here, a spontaneous re-innervation is unlikely to occur [2].

> **Acute para- and postinfectious neuropathies** are most frequently observed in classic peripheral facial nerve paresis (idiopathic, or infectious through Borrelia burgdorferi and varicella-zoster virus), and in a generalized form as demyelinating or axonal Guillain– Barré syndrome (GBS). These are either caused by a direct invasion by the pathogen into nerves and the spinal ganglia accompanied by inflammatory infiltrates (i.e., herpes zoster, herpes simplex, lepromatous and tubercular lepra, various parasitic agents), vasculitis disorders (borreliosis, early symptomatic HIV infection), demyelination of Schwann cells (Guillain–Barré syndrome, diphtheria), or blocking of axonal transmission by antiganglioside antibodies (axonal forms of GBS: acute motor axonal neuropathy (AMAN), acute motor-sensory axonal neuropathy (AMSAN)).

> **Nongenetically caused metabolic polyneuropathies** can affect children and adolescents, especially in conjunction with chronic kidney failure and diabetes mellitus [3,4]. Neuropathies caused by vitamin deficiencies occur through malnutrition, resorption disorders, and insufficient parenteral feeding via a shortage of vitamin B complex (B1, -2, -6, -12)

and vitamin E [5]. The cause of critical illness-neuromyopathy occasionally diagnosed in ICU patients undergoing respirator treatment is unclear [6].

**Toxic neuropathies** can be caused by medical drugs, heavy metals, organic solvents, and organic phosphoric acid esters. Their pathophysiologies are generally associated with axonal damage affecting various cellular mechanisms [5,7].

**Secondary neuropathies** in association with collagenoses or vasculitis syndromes are extremely rare in children and adolescents. They usually occur as multiplex mononeuropathies and are pathologically characterised by segmental perivascular infiltrates and axonal lesions.

**The non-syndromic hereditary neuropathies** are clinically, genetically, pathologically, and electrophysiologically heterogeneous. The largest group among them, formerly called HMSN, is now classified according to the OMIM system and in reference to the neurologists who first described them as Charcot–Marie–Tooth (CMT) neuropathies, depending on their genetic causes. Neuropathologically speaking, they are differentiated according to what is primarily damaged—namely the axon, or myelin sheath [8]. The hereditary sensory-autonomic neuropathies (HSN) and hereditary motor neuropathies (HMN) could be regarded as CMT-related diseases with some similar features.

In addition to these isolated hereditary motor-sensory neuropathies, there is a broad range of disorders of the peripheral nerves in **complex neurometabolic and neurodegenerative diseases** revealing mainly CNS symptoms [9–11]. Here, usually the peripheral neuropathy manifests as a minor symptom, it is rarely the primary manifestation leading to diagnosis. Specific disease-related biochemical or histopathological findings are frequently associated with these diseases.

#### **4. The Diagnostic Methods**

#### *4.1. History and Clinical Evaluation*

**Recommendation 1:** Should a neuropathy be suspected, we recommend that a thorough patient history be taken, including the family's history. In addition to their known medical problems, information on previous infections and toxin exposures should be acquired.

Strength of consensus: strong (10/10)

**Commentary:** The patient history should incorporate the initial symptoms and the course, and exposure to potentially causative factors in the patient's past and family history. Early symptoms to ask for include neonatal and infantile muscular hypotonia and problems with sucking and swallowing, delayed motor development and walking, clumsiness and poor co-ordination compared to peers.

**Recommendation 2:** We recommend that a clinical examination include testing the patient's skin and muscle trophics, their strength or paresis grade, and their reflex and sensory status.

Strength of consensus: strong (10/10)

**Commentary:** Peripheral neuropathies usually manifest as muscle weakness, loss of deep tendon reflexes, and distal muscle atrophy. Foot deformities and other contractures result from muscular imbalance and frequently manifest as pes cavus and in severe cases as equinus or club foot. Typical neurological symptoms include an inability to walk on heels, steppage gait, and abnormal co-ordination in walking and manipulation (doing buttons, peg-in-hole tests). Sensory anomalies may also appear whereby the function of the large sensory fibres (touch, deep sensitivity) is usually more strongly affected than small-fibre function (pain and temperature sensations). Sensory ataxia with a positive Romberg sign may become apparent. Autonomous skin disorders (coldness, hypohidrosis, hyperhidrosis) as well as the regulation disorders caused by impaired autonomic nerve function (i.e., bladder emptying problems) can occur.

**Recommendation 3**: We recommend that the indication for more extensive diagnostics (electrophysiology, imaging, CSF diagnostics, lab work-up, toxicology, molecular genetics, biopsy) be determined depending strongly on the patient's medical history and clinical findings.

Strength of consensus: strong (10/10)

#### *4.2. Electrophysiological Diagnostics*

Clinical neurophysiology plays a key role in diagnosing neuropathies [12]. The most essential aspects to investigate when neuropathy is suspected are:

	- Any signs of florid denervation or re-innervation
	- Signs of any additional involvement of CNS structures

It is essential that the diagnostic methods applied can be expected to yield information which extends beyond the patient's clinical findings. To spare children from unnecessary discomfort and worry, clinical and electrophysiological diagnostics must go hand in hand, and the physician needs to have the expertise and experience to ensure this [13,14]. Electrophysiological examinations, especially needle electromyography, are considered an invasive procedure. The examiner is assumed to possess not just knowledge of the age-specific normal values [15], but also a familiarity with the methods and psychological approaches young patients require. On suspicion of a dominantly inherited CMT it can make sense to investigate the parents instead of their young child.

**Recommendation 4:** We recommend that the electrophysiological diagnostics include motor and sensory neurographies. We suggest that the exam range be oriented along the concrete issues at hand and the patient's tolerance level.

Strength of consensus: strong (10/10)

**Commentary:** Motor and sensory neurographies help identify the neuropathies' fundamental pathomechanism, namely whether it is primarily demyelinating or axonal (Table 2). However, with some neuropathies, it is not possible to definitively differentiate between these (mixed types, intermediate types). If a compressed nerve or other circumscribed lesion is suspected, the clinician can attempt to localise the lesion by stimulating the nerves at different anatomical points along the nerve course (latency or amplitude jump, slowing in nerve conduction velocity (NCV) in the affected section, delayed distal motor latency). Most neuropathies in childhood also reveal obvious abnormalities in sensory neurography, for example, reduced amplitude in the sensory nerve action potentials (SNAP) and possibly a slower NCV. A normal sensory neurography in a patient with purely motor symptoms should make the physician consider dHMN or DSMA.

**Table 2.** Electrophysiological criteria for demyelinating or axonal damage to peripheral nerves.


NCV: nerve conduction velocity; CMAP: compound muscle action potential.

**Recommendation 5:** We suggest that the patient undergo electromyography when seeking signs of denervation as evidence for an acute or chronic axonal neuropathy, or when there is a suspicion of an accompanying or alternative myopathy. Strength of consensus: strong (10/10)

9

**Commentary:** An EMG is advisable when clarifying a neuropathy diagnosis to detect axonal lesions, i.e., for a case of acute axonal GBS, a CMT with a normal NCV, and to differentially distinguish a neuropathy from distal spinal muscle atrophy or myopathy (Table 2). This enables an assessment of pathologic spontaneous activity (positive sharp waves, fibrillations), the discharge and recruiting pattern of motor units, and the configuration of motor unit potentials (MUP). There is also an indication when seeking for signs of re-innervation, especially following traumatic neuropathy [2]. In addition, myotonic and neuromyotonic discharges can be useful in detecting diseases with muscular or nerve hyperexcitability.

**Recommendation 6:** We suggest that visual and/or acoustically evoked potentials be examined in case of a systemic disease involving the peripheral and central nervous system. Strength of consensus: strong (10/10)

**Commentary:** The examination of visual and auditory evoked potentials enables us to assess the functionality of these specific peripheral and central sensory pathways. However, the investigation of somatosensory-evoked potentials (SEP) often fails or is inconclusive because of peripheral nerve damage. Nonetheless, with a high number of stimulations, a central summation effect can sometimes result in a recordable SEP.

#### *4.3. Sensory and Vegetative Functional Diagnostics*

**Recommendation 7:** Should isolated small-fibre neuropathy be suspected, quantitative sensory testing can be performed.

#### **Strength of consensus:** consensus (8/10)

**Commentary:** Neuropathies affecting the thin or non-myelinated (Aδ- and C-) fibres necessary for pain and temperature sensation (small-fibre neuropathy) also occur in childhood and adolescence. However, they are often detected at a late stage or not at all because of their vague or uncharacteristic clinical symptoms such as pain, fatigue, and nausea. Auto-immune disorders, adolescent-onset Morbus Fabry and familial amyloid polyneuropathy (FAP), or other hereditary sensory neuropathies are known causes [16,17]. While routine electroneurography does not usually help diagnostically, the functional disturbance can be detected via quantitative sensory testing (QST). QST has been validated for children aged 6 years and beyond, but it is currently available mainly in pain clinics [3]. However, to objectively diagnose a small-fibre neuropathy or FAP, a skin biopsy and genetic testing is ultimately required [16,17] (see also neuropathology).

**Recommendation 8:** In case of suspected autonomic neuropathy, examinations to detect an autonomic function disorder can be performed (i.e., heart rate variability, sympathetic skin reactions, and the tilt-table test).

#### **Strength of consensus**: strong (10/10)

**Commentary:** The autonomic nervous system should be examined, for example in case of severe GBS and hereditary autonomous neuropathies to estimate the risk of heart arrhythmias and/or cardiac arrest. This examination is available in paediatric and adult cardiology departments [18].

#### *4.4. Imaging Diagnostics*

**Recommendation 9:** We recommend imaging procedures (ultrasound and MRI) for cases of local or regional neuropathies without a definitive aetiological explanation; imaging methods could reveal possible therapy-relevant lesions (nerve tumour, nerve entrapment syndrome, focal inflammation).

Strength of consensus: strong (10/10)

**Commentary:** Clinical neurological examinations are now enhanced by imaging technologies. The discovery of a causative tumour or compression enables surgery to de-compress the nerve.

**Recommendation 10:** In cases of polyneuropathies and diffuse neuropathies with doubtful clinical-electrophysiological findings and an indication to rule out an intraspinal or radicular tumour or prolapsed disc, we recommend MRI imaging or ultrasound.

Strength of consensus: strong (10/10)

**Recommendation 11:** Imaging via spinal MRI or ultrasound of the proximal nerves may also be performed to detect anomalies typical of inflammatory diseases (GBS, CIDP) in case the patient's findings so far have been inconclusive.

Strength of consensus: strong (10/10)

**Commentary:** In patients with Guillain–Barré syndrome, CIDP, and other inflammatory radiculopathies, the spinal MRI often reveals thickening and contrast-medium enhancement in the spinal and cranial nerve roots [19–21]. Examinations of circumscribed and diffuse lesions and diseases of the nerve plexus and peripheral nerves via highresolution nerve MRI or sonography are highly interesting [22–24]. However, as with electrophysiology, sonographic examinations of the peripheral, proximal, and cerebral nerves require a great deal of expertise and experience and are only available in few paediatric neurology and/or neuromuscular expert centres, making a general recommendation in these guidelines premature at this time.

#### *4.5. Laboratory and Other Paraclinical Diagnostics*

Laboratory parameters (Table 3) enable us to clarify secondary neuropathies induced by primary internal diseases (liver or kidney diseases, diabetes, collagenoses). If there are hints of a disease caused by a vitamin deficiency, the corresponding analyses should be undertaken, as in case a neurometabolic disorder in suspected [11].


**Table 3.** Laboratory tests for clinically and electrophysiologically suspicious cases.

#### *4.6. Neuropathological Diagnostics*

**Recommendation 12:** We recommend performing a nerve biopsy in case the diagnosis of a severe or progressing polyneuropathy is not otherwise possible, that is, via less invasive methods, and provided a firm diagnosis and therapy can be the consequence. This applies mainly to patients suspected of having vasculitis.

Strength of consensus: strong (10/10)

**Commentary:** Determining the indication for a nerve biopsy cannot be taken lightly; it requires great caution and differentiation. It is particularly significant for differential diagnostic workup when neuropathies are being considered that are not hereditary and can be effectively treated. In these categories belong infections such as vasculitis and perineuritis, as well as atypically presenting inflammatory neuropathies (chronic inflammatory

demyelinating or axonal neuropathy), nerves compromised by a lymphoma, and amyloid neuropathy [25,26]. Histological tests are generally done in the sensory sural nerve and require a compression-free section of the nerve. A part thereof is fixed in formalin for paraffin histology including Congo red or thioflavin-staining and immunohistochemistry; the other is fixed in glutaraldehyde (embedding in synthetic resin for semi-thin section/toluidine blue-staining and possible electron microscopy) [27]. Nerve biopsies should only be carried out and analysed in highly specialised centres; collecting adequate tissue from infants is extremely difficult and requires a very experienced surgeon [28].

Suspected inherited sensory neuropathy or small-fibre neuropathy (SFN) calls for a **skin punch biopsy** including immunohistochemistry of epidermal and dermal nerve fibres with staining for Protein Gene Product 9.5 (PGP9.5); it also allows the study of dermal myelinated fibres, autonomic innervation (sweat glands, arrector pili muscle, arteriovenous anastomosis), and mechanoreceptors. It is a less invasive procedure used to obtain diagnosis, but it also requires expertise and the knowledge of norm values appropriate to the given age group [17,29]. Adolescent-onset FAP can be suspected by the presentation of amyloid deposits but needs genetic testing of the TTR gene for confirmation [16].

#### *4.7. Genetic Diagnostics*

**Recommendation** 13: In the case of suspected hereditary neuropathy we recommend molecular genetic diagnostics which include different methods depending on the patient's clinical findings and family's medical history.

Strength of consensus: strong (10/10)

**Comment:** Over 100 genes have been reported as being responsible for causing CMT neuropathies. The diagnostic algorithm depends on multiple factors of the presenting patient. If the family is known to carry a specific mutation, that can be verified in the patient via MLPA or Sanger sequencing. If CMT neuropathy is suspected without a known mutation, the first diagnostic step should be to identify the PMP22 gene's copy number (usually via MLPA), especially in the case of demyelinating polyneuropathy. If that result is inconspicuous, massive parallel sequencing (next-generation sequencing (NGS)) is currently usually carried out, as it enables the most rapid and cost-effective analysis of many genes [8].

The **Gene Diagnostics Law** (GenDG) has been in effect in Germany since 2010. It mandates that special measures be taken to ensure that patients are well-informed before consenting to such tests. While any physician can schedule and carry out diagnostic examinations in patients with symptoms, the use of predictive genetic tests in healthy persons who carry a risk or in individuals with a possible genetic predisposition must be preceded by genetic counselling, which must only be conducted by certified physicians qualified to engage in genetic counselling and consultation. The genetic counselling must also include the information that molecular analysis may bring to light additional findings from genome diagnostics raising entirely different issues such as a hereditary cancer risk or risk for other neurological diseases.

#### **5. Differential Diagnosis of Acquired and Hereditary Neuropathies in Children and Adolescents**

#### *5.1. Nerve Injuries*

**Recommendation 14:** If peripheral nerve injury is suspected, we recommend a clinical examination to locate the affected nerve and lesion site, and to supplement this if needed by a neurophysiological investigation.

Strength of consensus: strong (10/10)

**Commentary:** Some injured or malfunctioning nerves are clinically practically impossible to identify; they can only be objectively assessed electrophysiologically (i.e., weakness in the palmar hand muscles with a failing pre-load in radial paresis) [2].

**Recommendation 15:** We recommend a liberal indication to perform imaging (especially ultrasound) diagnostics in patients with a nerve lesion, particularly when operative

treatment is likely necessary (i.e., neuronotmesis, nerve compression in a fracture gap, compression by a haematoma or tumour).

Strength of consensus: strong (10/10)

**Commentary:** Lesions affecting individual nerves or nerve plexus are often caused by typical accidents or traumas. Their clinical symptoms depend on the function of the affected nerve (motor-sensory mixed) and on the lesion's location. An entirely transected nerve leads to paralysis of the innervated musculature, loss of sensation, and when the N. medianus and tibialis are transected, to the loss of sweat secretion in that region. Should the patient reveal some remaining function, the likelihood of continuity and spontaneous recuperation rises. Table 4 illustrates the main postnatal lesions and their motor and sensory symptoms as well as their most frequent causes.

**Table 4.** Clinical appearance and aetiology of peripheral nerve lesions.


**Recommendation 16:** Electrophysiological tests can be done repeatedly to follow up patients and keep track of denervation and re-innervation processes, thus enabling a more accurate prognosis after a nerve lesion.

Strength of consensus: strong (10/10)

**Commentary:** A nerve's continuity can be confirmed via electrophysiological tests in patients suffering from total clinical paralysis following a nerve lesion. For patients who fail to fully recover after a nerve trauma, we suggest electrophysiological controls at 3-month intervals to evaluate the extent and direction of re-innervation, as they yield such information earlier than clinical examinations can [2].

#### *5.2. Mononeuritis, Mononeuritis Multiplex*

**Recommendation 17:** We recommend brain MR imaging in case of a suspected nonidiopathic facial nerve palsy; this is especially recommended in case of multiple cranial nerve lesions.

#### Strength of consensus: strong (10/10)

**Recommendation 18:** We recommend that infectiological blood tests be run for clinically suspected cases of borreliosis to confirm the aetiology; we suggest a lumbar puncture and cell count as well as serological tests in CSF be carried out to diagnose CNS involvement.

Strength of consensus: consensus (9/10)

**Commentary:** A peripheral or nuclear lesion of the facial nerve leads to paralysis in the mimic muscles innervated by all three of its branches. In contrast, a lesion in the corticobulbar tract leaves the function of the frontal branch intact thanks to bilateral cortical representation. Depending on where it is located, a nerve lesion along the facial canal in the base of the skull can cause lacrimal secretion to fail, as well as a loss the stapedius-reflex with hyperakusis and a loss of taste sensation on the affected side. Isolated facial paresis in childhood is usually idiopathic and inflammatory (Bell's palsy). However, during the summer and autumn, cases of facial paresis are frequently caused by neuroborreliosis. It is often accompanied by minor symptoms of meningeal irritation; mononuclear CSF pleocytosis is detected in over 90% of such cases. The presence of pleocytosis and elevated Borrelia antibody titers in the CSF are required as evidence of neurological involvement [30]. Other causes of facial nerve palsy are zoster oticus, otitis media, petrosal bone fractures, and tumours in the brain stem and cerebello-pontine angle [31].

**Further infectious forms of neuritis** are manifested in the context of specific infections (borreliosis, zoster, diphtheria, leprosy). These can strongly determine each disease's presentation, or go largely undetected as secondary phenomena. Their clinical symptoms are focal or multifocal, and cranial nerves are often affected. Cases of symmetric polyneuritis are seldom: in that case, it can be difficult to differentiate these from a post-infectious Guillain–Barré syndrome.

**Recommendation 19:** In cases of suspected vasculitis neuropathy, we recommend a biopsy if a firm diagnosis has proven impossible with other less invasive methods (for example, to detect typical antibodies).

Strength of consensus: strong (10/10)

**Commentary:** Cranial or spinal neuropathies or mononeuritis multiplex can appear in conjunction with different inflammatory systemic diseases. They can occur in lupus erythematodes, polyarteriitis nodosa, granulomatosis associated with polyangiitis Wegener, eosinophilic granulomatosis associated with polyangiitis Churg–Strauss, Boeck' disease, Schönlein–Henoch syndrome, inflammatory intestinal disorders, and other autoimmune diseases [32,33]. Guillain–Barré syndrome may also be present in conjunction with these illnesses, an important factor to consider in terms of the different therapeutic consequences.

#### *5.3. Guillain–Barré Syndrome (GBS) and Chronic Inflammatory Demyelinating Polyneuropathy (CIDP)*

We recommend consulting the corresponding S3 guidelines for the diagnosis and treatment of acute **GBS**, whose progressive phase is limited to 4 weeks [34]. CIDP has to be assumed if a patient exhibits a longer protracted, progressing, or fluctuating disease course.

**Recommendation 20:** In the case of prolonged demyelinating polyneuropathy revealing an obviously fluctuating or progressing course we recommend diagnostics for a suspected CIDP (including a CSF protein and cell count, possibly also a spinal MRI); we also recommend therapy attempts involving intravenous immunoglobulin (IvIG), plasmapheresis, or corticosteroids for inconclusive cases.

Strength of consensus: strong (10/10)

**Commentary:** CIDP can affect all age groups. Unlike acute GBS, CIDP exhibits a chronic, continuous or stepwise progressing or relapsing–remitting course. Paediatric diagnostic criteria for CIDP mandate a progressing period lasting at least 4 weeks. However,

up to 20% begin as acute GBS (aCIDP) and move from there to a chronic or relapsing course [35,36]. The clinical symptoms consist of motor and sensory impairments, where the impairment of just one function is more seldom. The neuropathy is usually symmetrically distributed, primarily in the distal legs. However, the symptoms can first appear in the arms and involve the neck muscles. The cranial nerves are often affected; however, respiratory insufficiency is less frequent than in patients with GBS. Additional diagnostic criteria comprise increased CSF protein in conjunction with a normal cell count, as well as an electrophysiological proof of multiple segmentally demyelinated nerves. Infectious, toxic, or metabolic neuropathies and a CNS process entailing a distinct sensory level and a paralysed sphincter must be ruled out. A high percentage of these patients lose the ability to walk unaided. The disease can last for months or many years [35–37]. Potentially effective therapies for CIDP are IvIG, plasmapheresis, or prednisolone of adequate duration (i.e., at least 3 months).

**Recommendation 21:** For patients with CIDP who are resistant to treatment with IvIG, plasmapheresis, and steroids, we recommend considering and testing for potential antibodies to paranodal proteins. We also recommend the patient to be re-examined for a potential hereditary aetiology (CMT). From adolescence onwards, familial amyloid polyneuropathy has to be considered, as it is frequently misrecognized as CIDP in the beginning [16].

Strength of consensus: strong (10/10)

**Commentary:** For children presenting a therapy-resistant and protracted course of a demyelinating polyneuropathy, we propose considering the possibility of a causation through antibodies to paranodal proteins like neurofascin-155; these "paranodopathies" may respond to rituximab even when the patient is IvIG-resistant [38–40]. Very slowly progressing CIDP is easily mistaken for a subacute course of **hypo- or demyelinating CMT**, and vice versa [41]. Increased CSF protein and excessive contrast medium absorption in the nerve roots on a spinal MRI are typical of CIDP, but neither is specific, and both can also accompany hereditary CMT neuropathies.

#### *5.4. Toxic Neuropathies*

**Recommendation 22:** We recommend that, in children also, the possibility of a toxic neuropathy always be considered. This can usually be ruled out by taking a careful patient history and ensuring the child undergoes a thorough clinical-electrophysiological examination. Specific laboratory investigations are rarely needed.

Strength of consensus: strong (10/10)

**Commentary:** A partial list of potential neurotoxic substances is found in Table 5.


**Table 5.** Toxic agents associated with polyneuropathy (selection).

*5.5. Hereditary Non-Syndromic Neuropathies in Children and Adolescents*

5.5.1. Hereditary Motor-Sensory Neuropathies (HMSN), Charcot–Marie–Tooth (CMT)-Neuropathy

#### **Clinical Presentations of CMT Neuropathies**

The clinical evidence of **classic CMT neuropathy** consists of symmetric weakness and atrophy in the distal leg muscles, weak deep tendon reflexes, and a neurogenic talipes

cavus. The disease's range of expression is extremely wide. These neuropathies usually become apparent in the **first** two decades of life. Most of their subtypes progress only slightly. Symptoms can spread to the thigh and hand musculature years later. Many of those affected exhibit few symptoms even in old age and are not detected until the family is screened. Yet, other individuals in the same family can present with early generalised weakness and suffer a very severe course.

The most frequent type of CMT is **CMT1**, the **demyelinating type.** It can be detected electrophysiologically, showing a homogeneously slowed motor nerve conduction velocity (NCV) of <38 m/s in the arm nerves (with the median nerve as reference nerve). The **axonal variant** of CMT neuropathy (**CMT2**) cannot be differentiated clinically from the demyelinating type; however, overlapping and mixed forms often make confirming a specific diagnosis very difficult. The situation is similar with **CMTX, the X-chromosomal dominant variant**, which is clinically usually apparent in males as CMT1, and in females often as axonal or mixed demyelinating-axonal neuropathy. **Additional symptoms** like hearing loss, optic atrophy, vocal cord paralysis, conspicuously rapid progression, atypical patterns in how the pareses are distributed, scoliosis, or renal insufficiency are potential signs of certain gene mutations [42,43].

**Congenital HMSN/CMT neuropathies** first becoming obvious in infancy are very rare, but nevertheless relevant for the paediatric neurologist. Because of their typical findings, these diseases used to be categorised as congenital hypomyelinating polyneuropathy and Déjerine–Sottas syndrome (previously CMT3) with demyelinating and hypertrophic neuropathy, pronounced sensory impairments, increased CSF protein, and a severe course. Molecular genetic evidence has shown that these are not independent genetic entities, but rather the clinically most severe manifestations of known phenotypically highly variable CMT mutations [44].

**Episodic neuropathies** are difficult to diagnose accurately because of their on–off phases. The most important form is autosomal-dominant **hereditary neuropathy with pressure palsies** (HNPP), with a prevalence of 7–16 individuals out of 100,000. This is characterised by recurring functional focal deficits in the peripheral nerves, especially at certain anatomically critical pressure points. It can develop over time into a chronic CMT1 or CMT2. Autosomal dominantly inherited **hereditary neuralgic amyotrophy** (HNA) is a further disease, which is characterised by burning shoulder pain and later, muscle atrophy. **Genetic Diagnostics for CMT Neuropathies**

It is much more successful to confirm a genetic diagnosis of CMT1 (50–80% of all CMT patients) than of CMT2 (10–30% of patients). According to larger studies following comparable protocols, we can genetically identify the four most frequent genes (PMP22, GJB1/Cx32, MPZ/P0, MFN2) in 40–60% of patients in whom an inherited neuropathy is suspected [8]. In patients with CMT1A, PMP22 duplication is detected in 50–70% of cases, GJB1/Cx32 mutations in 9–18%, and MPZ/P0 mutations in 3–10%. The study results are more variable in patients with CMT2: GJB1/Cx32 in 7–19%, MFN2 in 2–20%, and MPZ/P0 in 1–6% of cases. The most frequent genes associated with autosomal-recessive CMT types are GDAP1 and SH3TC2. Mutations in GDAP1 are the most prevalent finding in autosomal dominant and recessive CMT2 cases in Spain and South Italy [45,46]. The HINT1 gene is especially prevalent in eastern Europe (Czech Republic); its clinical presentation is that of an axonal neuropathy accompanied by neuromyotony [8].

**Recommendation 23:** We recommend molecular-genetic diagnostics in case of a suspected demyelinating CMT neuropathy, starting by determining the number of PMP22 copies. This can also be done as the initial diagnostic step in case of an axonal CMT neuropathy.

Strength of consensus: consensus (9/10)

**Commentary:** Quantitative analysis of the PMP22 gene via MLPA to detect deletions and duplications is a rapid, reliable test method, while these larger duplications and deletions are not detected via NGS and Sanger sequencing. PMP22 duplications are identified in 50–70% of CMT1 patients. PMP22 deletions seldom also manifest as axonal CMT: CMT2 was clinically diagnosed in 1% of a series of 334 patients with PMP22 deletions. This PMP22 deletion, however, amounted to 5.3% of the group with genetically proven CMT2 (n = 57), making it the fourth most frequent CMT2 aetiology [47].

**Recommendation 24:** Should an HNPP be suspected, we recommend (once a PMP22 deletion has been ruled out) analyses seeking intragenic PMP22 mutations.

Strength of consensus: strong (10/10)

**Commentary:** HNPP is characterised by mainly heterozygotic deletions of the PMP22 gene; loss-of-function point mutations in the PMP22 gene are found more seldom.

**Recommendation 25:** After ruling out PMP22 duplication or deletion, we recommend performing massive parallel sequencing of gene candidates when CMT neuropathy is suspected.

Strength of consensus: strong (10/10)

**Commentary:** It is expected that advances in genetic testing will identify more and more genes in which mutations lead to ultra-rare hereditary conditions including neuropathies, and that the frequency of causative mutations will be better defined via highthroughput technologies which enable numerous genes to be examined simultaneously. The most up-to-date information on the genes associated with neuropathies is found in the Online Mendelian Inheritance in Man data base (OMIM, www.ncbi.nlm.nih.gov/omim (accessed on 6 August 2021)). The US National Institutes of Health/NIH also provide an overview of genetic diagnostics for HMSN/CMT in their gene reviews (http://www.ncbi. nlm.nih.gov/books/NBK1358/ (accessed on 6 August 2021)).

5.5.2. Distal Hereditary Motor Neuropathies (dHMN)

This is a rare and genetically heterogeneous group of diseases characterised by exclusively motor symptoms affecting the distal muscle groups. Muscle atrophies and pareses primarily affect the lower extremities, but there are also forms more likely to affect the arm and hand muscles. There is some clinical and genetic overlap with distal spinal muscle atrophies that by electrophysiological and histological definition reveal no pathology in the peripheral nerves and exhibit neurogenic re-organisation in the affected musculature. dHMN can also overlap with spastic paraplegia (e.g., BSCL2). Thus, clinicians should pay attention to pyramidal signs and eventually perform motor evoked potentials.

**Recommendation 26:** We recommend performing massive parallel sequencing of gene candidates if a dHMN is suspected. PMP22 copies can be quantified beforehand.

Strength of consensus: consensus (9/10)

**Commentary:** The genetic clarification of dHMN/DSMA has improved greatly through high-throughput technologies in the last few years, and now lies in the range of 30–40% [48]. Because of dHMN's clinical and genetic overlapping with axonal CMT neuropathy, the genetic diagnostics are usually done simultaneously (Figures 1 and 2).

5.5.3. Hereditary Sensory and Autonomous Neuropathies (HSAN, HSN)

This group of extremely rare hereditary polyneuropathies is referred to as HSAN or HSN. It is primarily characterised by distally pronounced sensory-functional impairments and autonomic symptoms, and much less by minor motor impairments. Its classification in five different types becomes that much harder to follow the faster and more specifically new genes are being identified [49,50].

The autosomal-dominantly inherited **HSAN I types** usually become apparent in the second decade of life and begin with pain and impaired temperature sensation. Patients later suffer from the loss of other sensory capacities and spontaneous pain. The loss of sensory innervation triggers trophic anomalies and ulcers on the hands and feet, and not seldom gives rise to osteomyelitis and osteolysis. Less important are autonomic functional disorders (like excessive perspiration). **HSAN II–V** is inherited autosomal-recessively and usually first manifests in infancy or childhood; **HSAN II** is characterised by painless injuries (through the loss of sensation) and acrodystrophy and joint degeneration. **HSAN III** is also known as Riley–Day syndrome or familial dysautonomy; it causes autonomicregulation impairments, vomiting, and primarily psychomotor retardation. **HSAN IV**'s

dominant features are generalised anhidrosis accompanied by episodic bouts of fever already during infancy (CIPA) together with the loss of pain sensations and cognitive deficits. Children with **HSAN V** reveal no cognitive deficits, but otherwise a clinical presentation resembling that of HSAN IV [49].

‐ **Figure 1.** VENN diagram showing the distribution of mutations among clinical and electrophysiological groups of hereditary neuropathies (CMT/HSAN/dHMN) [8]. ‐

**Figure 2.** Algorithm to perform a genetic diagnosis of hereditary neuropathies [8].

**Recommendation 27:** Molecular genetic diagnostics via massive parallel sequencing of gene candidates is recommended to clarify cases of a suspected HSN/HSAN (see Figure 1).

Strength of consensus: strong (10/10)

‐

‐ ‐ ‐ **Commentary:** According to the very latest evidence, the responsible gene defects are only detected in 10–20% of families affected by HSAN/HSN [49,50]. As of the time these guidelines were compiled, no large investigations had been conducted targeting multi-gene panels or examination sequencing in association with this very rare group of diseases.

‐

‐

#### 5.5.4. Genetic Diagnostic Algorithms

To decide which genes should be analysed to clarify the aetiology of a suspected hereditary neuropathy, it is first important to clinically specify the disease as CMT neuropathy, episodic neuropathy, purely motor dHMN, purely sensory HSN/HSAN, or neuropathy in the context of a multi-system disease (Figure 2).

A quantitative procedure (MLPA) is initially recommended in the case of suspected demyelinating CMT neuropathy or HNPP to identify the frequent PMP22 duplication or deletions (Figure 2). While this diagnostic approach will fail in case of HSAN, for axonal CMT or dHMN it can be performed, although with less expectation of a positive result than in CMT1/HNPP. If the MLPA finding is negative, massive parallel sequencing of gene candidates is nowadays standard for nearly all hereditary neuropathies.

As the four most frequent genes (PMP22, GJB1/Cx32, MPZ/P0, MFN2) are detected in 90–95% of positively genetically screened patients belonging to cohorts of mainly middle-European heritage [8], clinicians are justified to first have these genes assessed after clinical and thorough formal genetic differentiation before expanding their genetic investigation to detect rare genes.

**Recommendation 28:** If an X-chromosomal CMT neuropathy is suspected, we recommend analysis of the GJB1 gene including coding regions, untranslated regions (UTRs), and promoter regions.

Strength of consensus: strong (10/10)

**Commentary:** 10–15% of patients with an X-chromosomal CMT reveal GJB1 mutations in non-coding regions that cannot be detected via high-throughput sequencing (Figure 2).

**As soon as a genetic diagnosis has been made, the affected family should be offered genetic counselling.**

#### *5.6. Peripheral Neuropathies Associated with Complex Genetic Diseases*

**Recommendation 29:** Should a systemic neurodegenerative or metabolic disorder be suspected in a patient with polyneuropathy, we recommend that primarily those diseases for which there are effective therapies be ruled out.

Strength of consensus: strong (10/10)

**Commentary:** Peripheral neuropathies can also occur in a complex array of symptoms caused by a large number of different neurometabolic and neurodegenerative syndromes. Thanks to the progress made recently in genetic diagnostics, their numbers have risen dramatically. In their systematic review, Rossor et al. [11] describe over 150 genetically defined diseases. Only a small proportion of these are manifested in childhood, and many have only been described in one or a handful of families. The neuropathy is often just one aspect of a given disease, although it can cause the main symptoms for some time during the disease's initial stage. For a given genetic disease, the neuropathy may present as an isolated symptom in adulthood if the mutation's effect is weak; but children whose gene is largely dysfunctional usually suffer from progressive CNS deterioration and early death. It is significant to note that effective therapies are available for some of these diseases provided they are diagnosed early enough, which is why they must not be diagnostically overlooked (Table 6).

**Table 6.** Neuropathies in the context of complex hereditary diseases amenable to treatment [11].



**Table 6.** *Cont*.

BMT: bone marrow transplantation; VLCFA: very long-chain fatty acids; gi: gastrointestinal.

A list of relevant systemic diseases in childhood is found in Table 7. The format of these guidelines did not permit the provision of more details. For more information, please see [9–11].

**Table 7.** Polyneuropathies in systemic neurological diseases with onset in childhood or adolescence.


EAOH: early-onset ataxia, with oculomotor apraxia and hypoalbuminemia; SCAR: spinocerebellar ataxia, recessive; SCA: spinocerebellar ataxia; AT: ataxia teleangiectasia; NARP: neuropathy, ataxia, and retinitis pigmentosa; ARSACS: autosomal recessive spastic ataxia of Charlevoix–Saguenay; SCAN: cerebellar ataxia and sensory-motor axonal neuropathy; SPG: spastic paraplegia; LBSL: leukoencephalopathy with brainstem and spinal cord involvement and lactate elevation; NBIA: neurodegeneration with brain iron accumulation; CDG: carbohydrate deficient glycoprotein disorder; CEDNIK: cerebral dysgenesis, neuropathy, ichthyosis, and palmoplantar keratoderma syndrome; ACPHD: ataxia, combined cerebellar and peripheral, with hearing loss and diabetes mellitus; PHARC: polyneuropathy, hearing loss, ataxia, retinitis pigmentosa, and cataract; MEDNIK: mental retardation, enteropathy, deafness, peripheral neuropathy, ichthyosis, and keratoderma; PTRH: infantile-onset multisystem neurological, endocrine, and pancreatic disease; **\* amenable to treatment.**

#### **6. Conclusions**

Diseases of the peripheral nerves with highly diverse aetiologies affect children and adolescents as well as adults. The spectrum of their causes in children differs strongly from that of adults. The differential diagnostic approach first requires fundamentally solid anatomical and neurophysiological knowledge and understanding, with thorough analyses of the disease course and family history, the disease's topographic distribution, and the quality of the patient's neurological symptoms. Supplemental examinations such as electrophysiology, laboratory workups, imaging, and CSF diagnostics follow a clinically driven hypothesis. That also applies to molecular genetic diagnostics, which require experience with and a command of the diagnostic algorithms due to the enormous genetic heterogeneity of hereditary neuropathies. These again will be expanded as the potential of broad-based panel or exome diagnostics is realized. To ensure effective and well-targeted therapy, a diagnosis as precise as possible is decisive—not just to alleviate nerve injuries, but also for patients with inflammatory and metabolic/malnutritional neuropathies. The causes of genetic neuropathies should not just be investigated in preparation for genetic counselling. Such knowledge is essential to be able to inform the patient about the disease's probable course and prognosis, and to enable them to participate in present and future clinical trials addressing pathologies and therapies.

**Author Contributions:** Conceptualization, R.K. and S.R.-S.; methodology, R.K.; software, not applicable; validation, not applicable; formal analysis, not applicable; investigation, not applicable; resources, not applicable; data curation, not applicable; writing—first draft preparation, R.K.; co-ordination of the Delphi process, R.K.; taking part in the Delphi discussion, all authors; writing—review and editing, all authors; visualization, not applicable; supervision, R.K.; project administration, R.K. and R.T.; funding acquisition, R.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** Translation of the manuscript was funded by the German Speaking Society of Neuropaediatrics.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** Publishing Society: Gesellschaft für Neuropädiatrie (R.K., R.T., M.B.). Participating Societies: Deutsche Gesellschaft für Kinder- und Jugendmedizin (R.T.), Deutsche Gesellschaft für Neurologie (B.S.), Deutsche Gesellschaft für Humangenetik (S.R.-S.), Deutsche Gesellschaft für Klinische Neurophysiologie (W.M.-F.), Deutsche Gesellschaft für Neuropathologie und Neuroanatomie (J.W.), Gesellschaft für Pädiatrische Radiologie (N.L., G.H.), Österreichische Gesellschaft für Kinderund Jugendmedizin (B.P.), Schweizerische Gesellschaft für Neuropädiatrie (G.M.S.), Schweizerische Gesellschaft für Pädiatrie (G.M.S.). We thank Carole Cürten Freiburg for editorial assistance and translation of the manuscript.

**Conflicts of Interest:** All authors are members of their respective societies. Otherwise, the authors declare no financial or non-financial conflict of interest in developing these guidelines.

#### **References**


**Agnieszka Cebula 1, \*, Maciej Cebula <sup>2</sup> and Ilona Kopyta 1**


**Abstract:** Ultrasonographic elastography is a relatively new imaging modality for the qualitative and quantitative assessments of tissue elasticity. While it has steadily gained use in adult clinical practice, including for liver diseases, breast cancer, thyroid pathologies, and muscle and tendon diseases, data on its paediatric application is still limited. Moreover, diagnosis of muscular diseases in children remains challenging. The gold standard methods, namely biopsy, electroneurography, and electromyography, are often limited owing to their invasive characteristics, possible contraindications, complications, and need for good cooperation, that is, a patient's ability to perform certain tasks during the examination while withstanding discomfort, which is a significant problem especially in younger or uncooperative children. Genetic testing, which has broad diagnostic possibilities, often entails a high cost, which limits its application. Thus, a non-invasive, objective, repeatable, and accessible tool is needed to aid in both the diagnosis and monitoring of muscle pathologies. We believe that elastography may prove to be such a method. The aim of this review was to present the current knowledge on the use of muscle elastography in the paediatric population and information on the limitations of elastography in relation to examination protocols and factors for consideration in everyday practice and future studies.

**Keywords:** ultrasonographic elastography; neuromuscular disease; muscle; children

#### **1. Introduction**

The diagnosis and monitoring of neuromuscular diseases remain a challenge despite the emerging role of genetic testing in this field. Vital limitations in diagnosis and monitoring are particularly relevant in the paediatric group owing to the high costs of genetic tests and invasiveness of gold standard tests (electromyography (EMG), electroneurography (ENG), and biopsy). Both ENG and EMG require good patient compliance (ability to simultaneously withstand discomfort and relax or contract muscles on demand), may lead to complications like most invasive procedures, and are often limited by the need to ensure patient safety and health-related contraindications. Thus, a new, non-invasive, affordable, and objective test is urgently needed. We believe that elastography may prove to be such a tool in the future, as it has already been applied in hepatological, endocrinal, and oncological diagnostics [1,2]. Although far more studies on muscle elastography have been conducted in the adult population than in children, results from the former cannot be simply applied to the latter. Thus, the aim of this study was to present existing elastography modalities, their limitations, and applications in paediatric muscle-related disorders.

#### *Technical Aspects of Elastography*

Elastography is an assessment method based on the elastic properties of soft tissues. A few modalities based on magnetic resonance (MR) and ultrasonography (US) already

**Citation:** Cebula, A.; Cebula, M.; Kopyta, I. Muscle Ultrasonographic Elastography in Children: Review of the Current Knowledge and Application. *Children* **2021**, *8*, 1042. https://doi.org/10.3390/ children8111042

Academic Editor: Rudolf Korinthenberg

Received: 20 September 2021 Accepted: 10 November 2021 Published: 12 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

exist. However, we limited the present study to US modalities, as they are easier to access in everyday clinical work, provide real-time metrics for most cases, and are less costly. Studies that compared elastography from MR and different US methods in children are scarce, and their results remain inconsistent [3–5].

The main US elastographic modalities are strain elastography (SE), acoustic radiation force impulse (ARFI), transient elastography (TE), and shear wave elastography (SWE). In general, the main distinctive differences are the method of stress application, detection of tissue deformation, and characteristics of gained data (qualitative vs. quantitative). A simplified classification of elastographic methods is presented on Figure 1. One of the most important impediments in elastographic research analysis is provider-dependant varieties, subdivision of modalities, and inconsistency in nomenclature.

SE is currently widely available, as it is offered by most leading manufacturers of US devices. The qualitative measurement is based on the amount of strain that is mechanically induced by either the mechanical pressure of the US probe or physiological processes such as blood vessel pulsation. The effect is presented as a colour map, with some vendors offering software for semi-quantitative analysis such as strain ratio (SR) calculation. Among the SE methods, a Hitachi-patented real-time elastography device has gained the most interest. It implements the extended combined autocorrelation method, an algorithm that correlates in both axial and lateral directions and produces an elasticity image in real time. Despite the implementation of various quality-control systems, the method is still heavily operator-dependant [1,6].

ARFI imaging is a Siemens-patented qualitative method based on the focused radiation force impulse produced by the US probe, whose displacement is evaluated at the set depth. The results are shown as a single image within the box. The attainability with systems by a single manufacturer is a valid availability limitation [1].

TE is the first quantitative method designed by Echosens for the evaluation of liver fibrosis and steatosis. A specifically designed piston induces a mechanical impulse, and US is used for beam-line average measurement of the resulting shear wave speed. Owing to its design, this method has not played a major role in muscular evaluation [1].

SWE is a group of methods that can be roughly divided into point and two-dimensional (2-D) SWEs, with further subdivision according to the method of force application. From a practical point of view, all these methods measure shear wave speed and offer quantitative results expressed in meters per second, which are subsequently converted to kilopascals by

using the Young modulus. One of the main limitations is the assumption of homogeneity of the wave propagation medium, which is usually not the case with muscles. Other limitations include the manufacturer-dependent method of induction and calculation of shear wave speed, which make a direct comparison of vendor results impossible [1,7,8]. SWE and ARFI may also be affected by the depth of examined tissues [7,9–11]. As in the paediatric population, tight muscle thickness was reported to be up to 45 mm. Thus, SWE might have a broader application in this age group [12]. In general the most commonly used devices are Aixplorer (Supersonic Imagine, Aix-en-Provence, France), which supports 2-D SWE, and Acuson (S2000/S3000; Siemens, Washington, DC, USA), which generates ARFI [13].

#### **2. Materials and Methods**

A thorough literature search on the PubMed database was performed using the MeSH terms "elasticity imaging techniques/methods", "child", and "muscle", and search words with the search operator AND ("elastography", "child", "muscle", and "children"). After the initial search of manuscripts from 2012 to 2020, 1329 studies were found. We limited the number of studies to 58 by manually checking articles and their abstracts. In addition, their bibliographies were analysed and checked. Studies that did not involve children, those that analysed only different muscular tissues (including studies on tendons only), those that had only the abstract available, duplicate articles, short reviews of the general use of elastography, those that involved only MR elastography, and those that were in languages other either Polish or English were excluded. Finally, 35 articles were included in this study. A summary of the process is presented in Figure 2.

*n* = 35

**Figure 2.** Flow diagram of the literature search.

The results presented were divided into two parts as follows: one is the technical and demographic aspects that influence the results, and the other is current knowledge on muscle elastography results in different clinical problems in children.

#### **3. Factors Influencing Elastography**

The following subparagraphs focus on factors affecting the elastography results that are not connected to the patient's disease. The sex, age, anthropometry, muscle stretching, tissue compression and operator-related reliability are discussed in turns.

#### *3.1. Differences Related to Patient Sex*

While multiple factors may influence muscular status, data regarding the association between patient sex and elastography results are too limited to form a clear conclusion. However, few studies have observed no sex-related differences between modalities in small children.

Brandenburg et al. found no significant sex-related difference in a study of SWE of the gastrocnemius muscle (GCM) in 20 healthy children aged 2 to 12 years. All the patients were in their prepubertal age [14]. No significant difference in ARFI was found in a study in 12 children with cerebral palsy (CP) aged 6 to 14 years and a study of SWE of the GCM in 86 patients from different age groups [15,16]. The latter involved 27 children (6–12 years) and 59 adults, but the authors did not assess the association of sex to muscles stiffness separately for the different age groups [15].

On the other hand, Koppenhaver et al. evaluated 130 adult patients to study factors that affect lumbar muscle stiffness on SWE. In this study, significant differences were found between men and women in all the muscles evaluated, with larger shear modules in the men. The authors estimated that the values for the male group were approximately 20% higher than those for the female group across all the examined lower back muscles in both the relaxed and contracted states [16]. Significant sex-related differences in ARFI and SWE findings were also described in other studies in adult populations [9,17,18]. One study of 42 healthy children found that there was significant difference between male and female's rectus femoris muscle but only at rest [19].

This sex-related disparity in elastography results between age groups and studies has raised the question of whether the observed changes might be dependent on maturity and the location and type of the examined muscles. More data are needed to clarify this.

#### *3.2. Age-Related Differences*

Increase in muscle stiffness with age was clearly described in some adult studies [9,17]. A similar tendency was observed in paediatric studies, with some minor deviations [14,15,20–22].

Liu et al. evaluated the GCMs of 86 healthy volunteers, divided into three age groups (paediatric: <16 years, middle-aged: 30–40 years, and old: >55 years). While no statistically significant differences were found in the plantar flexion (PF) and neutral positions of the feet, a significant increase in muscle stiffness in dorsiflexion (at all angles of the ankle from 10◦ to 30◦ ) was observed with age [15]. In a study by Wenz et al. that compared the SE results of the upper and lower limb muscles in young adults (22 patients, aged 20–30 years) and children (21 patients, aged 2–12 years), significant differences were found between the two groups [20]. Brandenburg et al. assessed muscle changes in 20 children (aged 2–12 years) and found that muscle stiffness increased with age but it did not reach statistical significance [14].

An interesting trend was described in a study of Achilles tendons in Turkish paediatric populations with and without CP. A significant difference in age was found, but while the SR increased with age in the CP group (*p* < 0.001), it showed an inverse correlation with age in the healthy group (*p* = 0.038). The authors concluded that changes in muscle stiffness not only occur due to ageing but also may show specific patterns of ageing in different diseases [22]. In a study in patients with Duchenne muscular dystrophy (DMD) aged 5 to 24 years, results were incoherent in that while a statistical correlation was found between age and GCM stiffness in the DMD group, no correlation was found in the healthy group or for other muscles in the DMD group [21].

#### *3.3. Differences Related to Anthropometry and Anisotropy*

No relationship was found between muscle stiffness and calf circumference in healthy subjects. A tendency for decreasing muscle stiffness in relation to increasing BMI was observed but did not reach statistical significance. Range of movement and foot dominance did not influence muscle stiffness [14]. Another study on children population presented incoherent results: while higher BMI was related with decrease of biceps brachii long elasticity immediately after exercise, but none at rest, for rectus femoris muscle it was the other way around. The authors found significant rise of elasticity with higher BMI but only at rest and not following exercise [19]. In the adult population in the study of Koppenhaver et al., a clear association was found between BMI, self-assessed activity level, and SWE results [16].

Gennison et al. studied the effect of a muscle's anisotropy, i.e., changes in mechanical properties along with the direction of measurement [23]. The longitudinal direction is recognized as the most relevant, as the muscle mechanical properties change with its lengthening and shortening [21,24]. The impact of the muscle anisotropy is still a subject of intensive research, with almost a hundred papers published annually. Yet, probably due to mentioned study results–authors of cited papers used longitudinal direction in their studies.

#### *3.4. Passive Muscle Stretching Influences Study Results*

Few studies focused on muscle stretching-related changes in muscle elasticity. Most of these studies assessed GCM changes as the effects of ankle joint dorsiflexion (DF), PF, and neutral position. Even though clear methodological differences (regarding the exact ankle angle measured, knee flexion, the study protocol used, and the tools used to ensure planned feet position) were present, the results were coherent; passive stiffness increased as a result of increased DF angle [14,15,25]. In a study of SWE of the anterior tibialis (AT) muscle and GCM in hemiplegic patients, Lee et al. confirmed that ankle angle had significant effects on both muscles. The study demonstrated a quadratic relationship between ankle angle and SW speed [26]. In a study by Brandenburg et al., muscle stiffness on SWE at 10◦ dorsiflexion was 4 times higher than at 20◦ PF [14]. Lacourpaille et al. proved that while muscle stiffness differed between patients with DMD and their healthy peers in 5 of 6 examined muscles during muscle stretching, this difference was present in only 3 muscles when no stretching was applied [21]. Thus, we may conclude that muscle stretching substantially changes muscle stiffness, especially in neuromuscular diseases, and not taking it into consideration may lead to errors in the study results and diagnosis.

Caliskan et al. focused on different aspects of muscle stretching. They studied whether the duration of passive muscle stretching affected muscle elasticity on SWE. Twenty male athletes aged 12 to 16 years were recruited. SWE was performed before and after 2 min (group 1) and 5 min (group 2) of passive stretching. While in the first group, no significant differences in pre-stretching results were found, significant reduction in muscle stiffness was observed in group 2 after 5 min. Even though the study population was small and limited to one sex, apart from the practical conclusions for sport medicine and physiotherapy, this study showed that the elastography examination protocol should carefully consider passive muscle stretching [25].

#### *3.5. Exercise and Effort's Effect on Muscle Elasticity*

Apart from mentioned above effect of stretching, effort and exercise lead to changes seen in elastography on their own. In a study of 40 paediatric patients, Berko et al. proved that results of strain elastography before and immediately after leading to fatigue exercise significantly differed. The post-exercise elasticity of both biceps brachii long and rectus femoris muscle was lower [19]. This effect must be taken into account while planning clinical use of muscle elastography.

#### *3.6. Influence of Tissue Compression on Elastography*

As muscle tissue reacts to compression and passive elongation, with possible exacerbation of the reaction in some neuromuscular diseases, the question of whether force applied by the US probe changes the elastography results remains valid. In most of the studies included in this analysis, researchers decided to minimise the effect by either performing the study with as little compression as possible or reducing the distortion of the

subcutaneous soft tissue with the use of a generous amount of US gel [4,14,18,27,28]. The latter solution might have an important limitation as seen in a study in 23 adult patients by Alfuraih et al. The authors compared SWEs of the vastus lateralis muscle with minimal pressure applied to the skin and with "standoff gel" applied at a minimum thickness of 5 mm. While no significant changes in mean shear wave speed was found between the groups, the reliability quantified by intraclass correlation coefficient (ICC) decreased from near-perfect agreement (ICC = 0.83) to the margin of substantial agreement (ICC = 0.62) in the gel method [7]. Further studies on the compression effect of the probe on elastography results are needed.

#### *3.7. Operator-Related Reliability*

A few studies assessed inter- and intra-operator reliability as moderate to excellent regardless of the muscle evaluated and modalities used (ICC, from >0.6 to >0.9) [24,29–31] A. While no relationship between BMI and elastography results in children were described by Linek et al., they found that the thickness of fat tissue above the lateral abdominal muscles influenced the reliability of the results of their study. Thinner fat layers (<5 mm on average) had positive influences on inter-rater reliability. In addition, the side of the examined muscle carried consequences as well. Worse inter- and intra-operator reliability results were observed when the examined muscle was on the patient's side opposite to the examiner. This again may be the result of the different pressures applied by the probe and thus again raises the question on methodology [29].

#### **4. Elastography in Different Muscle Disorders**

The following subparagraphs focus on the elastography results in relation to the patient's muscles disease. The muscular dystrophies, other myopathies, cerebral palsy and its treatment are discussed in turns.

#### *4.1. Muscle Elastography in Muscular Dystrophies*

As research on treatment strategies for myopathies has been progressing over the recent years, the need for treatment monitoring tools is also increasing. Few studies have presented differences between healthy peers and patients with DMD. Lacourpaille, Lilian, et al. compared results from different muscles in a healthy group (*n* = 13) and a DMD group (*n* = 14) of patients aged 5–24 years. While significant differences in muscle stiffness were found in the AT, GCM, vastus lateralis, biceps, and triceps brachii, the difference in muscle stiffness of the abductor digiti brevis minimi did not reach statistical significance. The highest difference was observed in the vastus lateralis, with the stiffness 134% higher in the DMD group [21]. Pichiecchio et al. compared the SWE and MRI results of the lower limb muscles (GCM, AT, rectus femoris, vastus lateralis and medialis, adductor magnus, and gluteus maximus) from 5 children with DMD with those from their age-matched healthy peers. Moderately higher muscle stiffness values were found in the DMD group. However, no significant correlation was found between the SWE and MRI results for fat replacement and muscle oedema on T1 and short inversion-time inversion-recovery (STIR) images. Although the study was limited by both the small number of participants and inclusion of a patient with a milder clinical presentation and mutation associated with Becker dystrophy, it presented the question of whether MRI and SWE analyse overlapping or different aspects of muscle diseases and if they can be used interchangeably. In addition, the patient with a clinical presentation of Becker dystrophy showed interesting results in that the changes in the SWE values were not accompanied by fat or STIR changes on MRI and clinical abnormalities in the patient examinations, which possibly preceded the latter [4]. The hypothesis that elastography might be useful as a screening tool thus remains to be proven.

Furthermore, by analysing changes in SWE results from patients with DMD over 12 months, Lacourpaille proved that elastography is a good candidate monitoring tool. They compared resting shear modules from the AT, GCM, biceps and triceps brachii,

and abductor digiti minimi muscles in 10 children with genetically confirmed DMD and 9 age-matched healthy peers. While no significant changes over time were found in the control group, the DMD group showed significant increases in the AT (75.1% ± 93.5%, *p* = 0.043), GCM (144.8% ± 180.6%, *p* = 0.050), and triceps brachii (35.5% ± 32.2%, *p* = 0.005). The biceps brachii and abductor digiti minimi muscle changes did not reach statistical significance [32].

#### *4.2. Other Myopathies*

Berko et al. designed a study to evaluate the usefulness and efficacy of SE for juvenile idiopathic inflammatory myopathies (JIIMs) and compared them with those of MRI. The authors recruited 18 patients aged 3 to 19 years, assessed the clinical activity of the disease, and performed both MRI of the pelvic region and thighs and strain elastography of the quadriceps muscles. The results were not favourable for SE. While the MRI results were related to disease activity (*p* = 0.012), the elastography results showed no association with either the MRI results or disease activity. Also no relationship was found between elastography and disease duration; thus, the results cannot be simply explained by the shorter disease time in the adults [5]. The question remains as to whether qualitative and less operator-dependent elastographic modalities would provide better results, as JIIMs are expected to affect muscle elasticity.

By contrast, Song et al. examined the SEs of patients with inflammatory myopathies, regardless of age. They proved that the affected muscles had higher strain rates and that SR correlated with the pathological scores of the biopsy samples [12]. However, as only one of the 17 patients was a child (an 11-year-old girl with juvenile dermatomyositis) and her SR was lower than those of the other participants, no conclusion for the paediatric population may be given from this study. This also implies the possibility that age influences elastography results.

#### *4.3. Cerebral Palsy*

CP is one of the most common disorders associated with secondary muscle changes in the paediatric population. Owing to the large number of patients, the patterns of changes found in affected muscles, and the treatments aimed at decreasing muscle spasticity, only a few studies on elastography in CP already exist.

Few studies focused on assessing the differences between muscles affected and those not affected by CP. Kwon at al compared strain and ARFI results from the GCM and soleus muscle in 15 patients with CP and 13 healthy peers, all aged <13 years. The GCM had greater muscle stiffness on SE and higher ARFI velocity in the CP group. While the soleus muscle had higher values on SE, its shear wave velocity was similar in both the CP and healthy groups. The SR (ratio of the GCM to the soleus muscle) was significantly higher in the CP group than in the healthy group. The authors concluded that the GCM had greater involvement in the motor deficits in CP [33]. In another study, Lee at al., based on a group of 7 children with hemiplegic CP and 1 post-stroke paediatric patient with similar clinical presentations, proved that the AT and GCM had significantly different SW speeds between the less and more affected sides. Patients with Gross Motor Function Classification System levels I and II, indicating no gross motor deficits, were compared. In the neutral joint position, the mean SW speed was 20% higher for the AT on the more affected side (3.86 m/s vs. 3.22 m/s, *p* = 0.03) and 14% higher for the GCM (5.04 m/s vs. 4/46 m/s, *p* = 0.024) [26]. Ozturk et al. compared Achilles tendon stiffness between CP patients (72 children, CP group) and their healthy peers (83 children, control group). The control group had lower SR than the CP group (1.7 ± 0.1 vs. 4.1 ± 0.8, *p* < 0.001) [22]. On the basis of the SWEs of the soleus muscle in 21 children with CP and 21 healthy peers, Vola et al. proved that muscle elasticity differed significantly between the hemiplegic and healthy patients. Higher Young modulus values were found in the CP group than in the healthy group (8.1 ± 2.3 kPa vs. 4.8 ± 1.7 kPa, *p* < 0.001) [34]. Similar results were presented for ARFI by Bilgici et al. In their study in 17 children with CP and 25 healthy

peers, they compared GCM elasticity and modified Ashworth scale (MAS) scores. The mean shear wave velocity was 3.17 ± 0.81 m/s in the CP group and 1.45 ± 0.25 m/s in the control group (*p* < 0.001) [35]. A recent study by Lallemant-Dudek et al. compared SWE results of GCM and biceps brachii long. Paper has some important limitations as the control group included patients with scoliosis while no information on possible muscle disease was given and in addition some of the patients from CP did receive botulin toxin (BoNT-A) treatment. Yet authors found that while there was no difference between groups when muscle was at rest, CP involved muscles differed from both–control group and uninvolved muscles in CP patients [36]. In other study, Mansouri et al. confirmed the relationship of the elastography results of the anterior tibialis muscle and GCM with gait abnormalities (step time and walking speed) [28]. Some studies confirmed that muscle stiffness assessed using elastography is related with clinical presentations assessed using MAS score [34,35]. While in others no such correlation was found [36].

To test the hypothesis that hyperactivity and spasticity of the hip abductors and flexors result in the hip displacement in CP, Doruk Analan et al. analysed the correlation between the Reimers hip migration index and the elasticity of the mentioned muscles on SWE. No significant correlation was found [37].

#### *4.4. CP Treatment Evaluation Using Elastography*

The use of elastography in monitoring botulin toxin A (BoNT-A) treatment in patients with CP is gaining attention. Studies that showed changes in muscle stiffness after BoNT-A administration into GCMs in children are presented in Table 1. In addition, some authors presented clear associations between post-BoNT-A changes in elastography result and clinical scale scores for spasticity (MAS and modified Tardieu scale [MTS]) [27,35,38]. By assessing post-BoNT-A elasticity changes of the anterior tibialis in addition to those of the gastrocnemius, Dag et al. proved that botulin treatment affects not only the muscle where BoNT-A was administrated but also the overall patient gait and related muscles. In some patients, paradoxical increases in shear wave speed were observed regardless of changes in MAS score, probably due to abnormal collagen content, injection failure, wrong injection site, insufficient dose, and so forth. In such cases, measurement of changes between the preand post-elastography values combined with US might prove to be a useful tool for further decision making regarding eventual treatment withdrawal [39]. In addition to mentioned studies that focused on the treatment effects at one month post BoNT-A administration, in a study in 9 children aged 2 to 9 years, Brandenburg et al. quantified the duration of treatment effect. They set three study visits up to 1 month before and 1 and 3 months after injection. While near-significant differences were found between the pre-BoNT-A administration and first post-treatment control, no significant difference was found between the pre- and 3-month post-BoNT-A values. The most significant difference was found between the 1 and 3-month post-BoNT-A values. Thus, the authors showed that the BoNT-A effect on muscle stiffness on SWE lasted <3 months post injection. SWE proved to be a reliable tool for individualised monitoring and planning of botulin treatment in patients with CP [40]. We can then conclude that elastography might be useful in the most effective and patient-tailored BoNT-A therapy by guiding administration planning, determining the treatment prognosis, allowing for objective treatment assessment (e.g., comparison between elastography-based MAS and MTS scores less subjectively and isolated muscle assessment rather than combined assessment of muscles, joints, tendons, surrounding tissues, and excitability changes), and guiding the choice of the best time interval between doses.


**Table 1.** Studies on the relationship between elastography and BoNT-A treatment in cerebral palsy.

ARFI, acoustic radiation force impulse; BoNT-A\*, botulin toxin A; PF, plantar flexion; RTS, real-time sonoelastography; SE, strain elastography; SWE, shear wave elastography; SWS, shear wave speed; yo, years old.

#### **5. Muscle Elastography in Other Diseases**

The following subparagraphs focus on the elastography results in relation to other diseases influencing muscle tissue. The chronic kidney diseases, gluteus muscle contracture, torticollis, Osgood-Schlatter disease, elbow injuries, musculoskeletal tumours are be discussed in turns.

#### *5.1. Chronic Kidney Diseases*

In their study, Bekci et al. assessed the possible use of ARFI in screening for muscle changes in chronic kidney diseases (CKDs). The reason for the muscle function loss in CKD is still not fully elucidated, but factors such as disease-related myopathy, muscle loss, and abnormal fat deposition are considered possible causes. The study population consisted of children aged 6 to 17 years, including 23 patients with CKD (11 girls) and 22 healthy peers (11 girls). The authors performed an elastographic evaluation of the elasticity of the rectus femoris muscle and handheld dynamometry (HHD) for evaluation of the maximal isometric strength of the knee extensors. The results showed that both muscle strength and elasticity were significantly decreased in the CKD group compared with the healthy volunteers. Whereas HHD has limited reliability, the authors concluded that elastographic techniques might prove feasible, affordable, and objective tools for treatment planning, monitoring, and screening for muscle changes [35].

#### *5.2. Gluteus Muscle Contracture*

Gluteus muscle contracture is a clinical syndrome most often found in the age group of 6–18 years, in relation to the above-mentioned pathological muscle changes, and is characterised by abnormal gait and hip movement limitations (mainly flexion and adduction). Diagnosis is often delayed and thus affects the prognosis, which is closely related to early treatment (surgery being the gold standard). In a small group of three patients, Guo et al. proved that measurement of muscle stiffness using SWE might be useful in the diagnosis. They speculated that SWE results may be related to the severity of the syndrome, making elastography a potentially useful tool for the assessment needed for treatment and prognosis [41].

#### *5.3. Torticollis*

Lee at al performed strain elastography of the sternocleidomastoid muscle and compared the results between infants with congenital torticollis and their healthy peers. The torticollis group had significantly lower muscle elasticity values [31]. The authors concluded that elastography may be a useful tool for monitoring and diagnosing torticollis especially in cases with subtle changes.

#### *5.4. Osgood-Schlatter Disease*

On the basis of real-time tissue elastography results of the rectus femoris muscles in 37 teenage male athletes, Enomoto et al. rejected the hypothesis that one of the factors that lead to Osgood-Schlatter disease is higher-than-normal muscle stiffness. No significant difference in quadriceps muscle stiffness was found between the OGD and healthy groups [2].

#### *5.5. Elbow Injuries Related to Sports*

In a study based on strain elastography results of the upper limb muscles in 197 baseball players aged 9 to 15 years, Saito et al. focused on addressing the question as to whether elbow injuries are related to pronator teres muscle (PTM) stiffness. The muscle group functions as a dynamic stabiliser against elbow valgus force. According to US results, the participants were divided into three groups, namely those with medial epicondylar fragmentation in the throwing arm, those with osteochondritis dissecans (OCD) of the humoral capitellum, and healthy peers. The elasticity of the pronator teres muscle was significantly higher in both injury groups than in the healthy group. In addition, the authors found that while only the muscle spasticity of the throwing arm was significantly higher in the OCD group, the PTMs on both sides were affected in the medial elbow injury group. The authors concluded that this may prove that medial elbow injury might be the result of muscle changes, not the other way around; thus, screening for muscle stiffness changes might help prevent the injury. Moreover, by comparing the strain results between the range of movement of the upper limb and those of the elbow and arm joints (with significant differences only for external rotation of the glenohumeral joint), the authors remarked that the changes in muscle spasticity observed on elastography might precede those observed in clinical examinations [30].

#### *5.6. Oncology: Musculoskeletal Tumours*

Timely detection and diagnosis of suspicious lesions are often the factors that lead to better treatment efficacy. However, the heterogeneity of pathological masses challenges all available techniques. Li et al. evaluated the usefulness of real-time 2-D SWE in distinguishing between benign and malignant musculoskeletal lesions by examining 115 tumours in 92 children and adults. Both quantitative (minimum, maximum, and mean elasticity in kilopascals) and qualitative (colour map sale) elastography results were analysed in comparison with the histopathological evaluation results. All the parameters were significantly different between the benign and malignant tumours (*p* < 0.05). By performing a multivariate regression analysis, the mean elasticity values were found to have strongest independent prediction for malignancy, with 71.4% accuracy, 66.7% sensitivity, and 85% specificity. In the same study, the authors also compared the diagnostic efficacy of US with those of both qualitative and quantitative SWEs. They found no significant differences in diagnostic efficacy, which was considered moderate for all the techniques. Assessment of lesion stiffness proved to be an important addition in morphological evaluations. One of the important limitation of this study was the exclusion of tumours exceeding the maximal region of interest of 4 × 6.5 cm [42].

#### **6. Study Limitations and Conclusions**

Since most existing studies have different study protocols in relation to the different positions of the analysed limb/muscles, which affected the acquired measurements; the stretching protocol; the non-homogenous groups with regard to age and sex; and the lack of standardisation of the pressure generated by the transducers on the patients' skin and muscles, the results of the studies included in our analysis were not comparable in most aspects.

Further studies are needed to develop normative values for different age groups that account for developmental changes, to characterise the influences of sex on the normative values, to standardise the test protocol, and to assess whether skin and tissue compression

significantly changes measured values. Moreover, further investigation into the probe position on muscles (perpendicular and distal/proximal positions), the influence of body temperature, muscle stretching, different muscles, and the relationships of stiffness values to clinical characteristics (e.g., hypotonia) and other diagnostic tools are needed. Nevertheless, application of this imaging modality is a promising direction for the diagnosis and monitoring of muscular diseases. Development of unified examination protocols and further objectification of muscle elastography may be an important step for better understanding, recognition, and monitoring of the different muscle pathologies in specific, non-easily cooperating group of small patients.

**Author Contributions:** Conceptualization, A.C. and M.C.; methodology, A.C.; software, M.C.; validation, A.C., M.C. and I.K.; formal analysis, I.K.; investigation, A.C.; resources, M.C.; writing—original draft preparation, A.C., M.C.; writing—review and editing, A.C., M.C., I.K.; visualization, A.C., M.C.; supervision, I.K.; project administration, M.C.; funding acquisition, M.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Etiology of Carpal Tunnel Syndrome in a Large Cohort of Children**

**Christina T. Rüsch 1,2,† , Ursula Knirsch 1,† , Daniel M. Weber 3 , Marianne Rohrbach 4 , André Eichenberger 5 , Jürg Lütschg <sup>2</sup> , Kirsten Weber 6 , Philip J. Broser <sup>2</sup> and Georg M. Stettner 1, \***


**Abstract:** (1) Background: Carpal tunnel syndrome (CTS), a compressive mononeuropathy of the median nerve at the wrist, is rare in childhood and occurs most frequently due to secondary causes. (2) Methods: Medical history, electrodiagnostic findings, and imaging data of patients with CTS from two pediatric neuromuscular centers were analyzed retrospectively. The etiology of CTS was investigated and compared with the literature. (3) Results: We report on a cohort of 38 CTS patients (*n* = 22 females, *n* = 29 bilateral, mean age at diagnosis 9.8 years). Electrodiagnostic studies of all patients revealed slowing of the antidromic sensory or orthodromic mixed nerve conduction velocities across the carpal tunnel or lack of the sensory nerve action potential and/or prolonged distal motor latencies. Median nerve ultrasound was diagnostic for CTS and confirmed tumorous and vascular malformations. Etiology was secondary in most patients (*n* = 29; 76%), and mucopolysaccharidosis was the most frequent underlying condition (*n* = 14; 37%). Idiopathic CTS was rare in this pediatric cohort (*n* = 9; 24%). (4) Conclusion: Since CTS in childhood is predominantly caused by an underlying disorder, a thorough evaluation and search for a causative condition is recommended in this age group.

**Keywords:** carpal tunnel syndrome; median nerve neuropathy; electrodiagnostic studies; neuromuscular ultrasound; mucopolysaccharidosis

#### **1. Introduction**

Carpal tunnel syndrome (CTS) is a compressive mononeuropathy of the median nerve at the wrist. In contrast to CTS in adult patients, the condition in childhood is rare, often manifests with atypical symptoms, and most frequently occurs secondarily due to other causes. In children, CTS was first described by Martin and Mass in 1958 [1], who reported on three children with recurrent episodes of hand pain. In 1989 Poilvach [2] carried out an extensive literature search and presented 52 cases of childhood CTS. He suggested the first etiopathological classification of the various underlying causes. Van Meir and De Smet [3,4] continued this work and performed a meta-analysis of 163 cases from 35 articles, mostly case reports or small case series.

**Citation:** Rüsch, C.T.; Knirsch, U.; Weber, D.M.; Rohrbach, M.; Eichenberger, A.; Lütschg, J.; Weber, K.; Broser, P.J.; Stettner, G.M. Etiology of Carpal Tunnel Syndrome in a Large Cohort of Children. *Children* **2021**, *8*, 624. https://doi.org/ 10.3390/children8080624

Academic Editor: Rudolf Korinthenberg

Received: 25 June 2021 Accepted: 21 July 2021 Published: 23 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The diagnosis of CTS in adults is primarily based on clinical symptoms and can be confirmed with electrodiagnostic studies [5]. In children, symptoms are often atypical, which reinforces the importance of technical investigations. Regardless of the etiology, isolated slowing of sensory or mixed nerve conduction velocity and/or prolongation of the distal motor latency (DML) of the median nerve across the carpal tunnel are electrophysiological hallmarks for CTS. Recently, neuromuscular ultrasound has been recognized as a valuable method for different neuromuscular conditions including entrapment neuropathies. This applies also to the evaluation of CTS [6]. Characteristics of median nerve ultrasound studies consist of an increase of both the cross-sectional area (CSA) at the wrist and the wristto-forearm ratio (WFR) [6]. For the majority of pediatric CTS cases, an underlying cause can be found, in particular hereditary metabolic conditions with mucopolysaccharidoses and mucolipidoses as the largest disease group, followed by congenital malformations, connectivopathies, endocrinopathies, and acquired lesions like malignancies or tumor-like and traumatic lesions [2,3,7–9].

The aim of this study was to investigate the etiology of CTS in a cohort from two tertiary pediatric neuromuscular centers in Switzerland (University Children's Hospital Zurich and Children's Hospital of Eastern Switzerland St. Gallen, Switzerland). We retrospectively analyzed the data of pediatric patients with CTS and evaluated diagnostic procedures and findings.

#### **2. Materials and Methods**

We retrospectively analyzed data of patients diagnosed with CTS in two tertiary pediatric neuromuscular centers in Switzerland (University Children's Hospital Zurich and Children's Hospital of Eastern Switzerland St. Gallen). Patients with an age below 18 years at diagnosis of CTS with characteristic electrophysiological findings were included in our study. The main focus of this study was to investigate the etiology of childhood CTS. Therefore, demographic data, medical history, manifesting symptoms, examination findings, underlying conditions, and proportion of etiologies were analyzed. For identification of CTS patients, the clinical information system of the two participating centers and registers of electrophysiological and surgical interventions were screened for the diagnosis of CTS. All patients who were diagnosed with CTS in the years 2005–2020 at the University Children's Hospital Zurich and 2016–2020 at the Children's Hospital of Eastern Switzerland, St. Gallen were included. All patients gave their consent to be included in our study.

For inclusion, all patients had to fulfill standard electrodiagnostic criteria for CTS. Since this is a retrospective work, different electrophysiological standard procedures established for the investigation of adults were performed [10]. Midpalm stimulation of the median and ulnar nerves and determination of latency differences between the orthodromic mixed nerve potentials at the wrist at a distance of 6–8 cm was preferred, because this method is least dependent on the small size of the hand in younger children, in which standard distal distances used in other electrodiagnostic approaches sometimes cannot be respected. Any latency difference, referred to as "palmdiff", of ≥0.4 ms was considered diagnostic [10]. Alternatively, fractioned antidromic sensory nerve conduction studies with stimulation at the wrist and midpalm and recording of sensory nerve action potentials on the second digit were performed. Slowing of the sensory nerve conduction velocity of ≥10 m/s across the carpal tunnel was considered diagnostic for CTS. In addition, distal motor latencies (DML) of the compound muscle action potential (CMAP) recorded from the abductor pollicis brevis muscle were obtained after median nerve stimulation at the wrist with a distance of 7.0 cm whenever possible. A DML of ≥4.1 ms was considered diagnostic [10]. Two different ENMG systems were used in our centers: Viking Monograph (Nicolet Biomedical Inc. Madison, WI, USA; used in Zurich) and System Plus (Micromed, Venice, Italy; used in St. Gallen).

In addition, ultrasound imaging data of the median nerve were analyzed if available. Median nerve ultrasound (US) imaging was done in 24% of the patients by a pediatric radiologist and/or pediatric neurologist trained in peripheral nerve US following standard

procedures [6]. The presence of structural changes was investigated along the median nerve. In addition, the median nerve cross-sectional area (CSA) at standard locations was measured and the wrist-to-forearm ratio (WFR) calculated with reference to age related nerve US normal values [11,12]. In both centers a Canon Aplio i800 (Canon Medical Systems, Tokyo, Japan) ultrasound imaging system, equipped with i33LX9, i24LX8, i18LX5, and i22LH8 was used.

The study was approved by the local ethics committee on 23 June 2020, and registered with the Swiss project database (BASEC 2020-01016). Written informed consent was obtained from the caregivers prior to inclusion of participants in the study.

#### **3. Results**

#### *3.1. Demographics*

We identified 38 patients (*n* = 22 females, *n* = 16 males) diagnosed with CTS in the two pediatric neuromuscular centers between 2005 and 2020 in Zurich and 2016 and 2020 in St. Gallen. The demographics of the patients are shown in Table 1. See Supplementary Table S1 for more detailed individual information.

**Table 1.** Demographics of childhood CTS patients.


#### *3.2. Etiology*

Lysosomal storage diseases (mucopolysaccharidosis and mucolipidosis) were the most frequent underlying conditions in our cohort (*n* = 15; 39%). Five patients (13%) showed CTS associated with a hereditary neuropathy (*n* = 3 probable hereditary neuropathy with liability to pressure palsy (HNPP) with positive family history, *n* = 1 Charcot-Marie-Tooth CMT type 1A, *n* = 1 associated with autosomal recessive spastic ataxia of Charlevoix-Saguenay (ARSACS)). In the three patients with a positive family history for HNPP the parents did not consent to genetic testing for their children. We identified CTS due to congenital malformations in four patients (11%). Two of them had been diagnosed with geleophysic dysplasia, one with the ultra-rare condition melorheostosis and one with a hemihypertrophia syndrome of unknown etiology. In two patients (5%), CTS occurred due to a benign tumor (*n* = 2 perineurioma, Figure 1), of which one perineurioma was associated with a PIK3CA gene mutation. A posttraumatic CTS was found in two patients (5%). One patient (3%) suffered from bilateral CTS associated with rheumatoid arthritis. Altogether, a secondary CTS etiology was confirmed in 29 patients (76%). In only nine patients (24%) was CTS considered idiopathic because of the absence of other explaining findings. See Table 2 for a summary of the CTS etiology.

**Figure 1.** Ultrasound of the median nerve in childhood CTS. (**A**,**B**) Idiopathic CTS with transverse sonogram of the median nerve at wrist (**A**) and forearm (**B**). The median nerve ultrasound investigation demonstrated a pathologically increased WFR of 2.3. (**C**–**F**) Intraneural perineurioma in two patients: Transverse sonograms of the median nerve at wrist (**C**) and forearm (**D**) of one patient, and transverse (**E**) and longitudinal (**F**) sonogram at the wrist of the second patient with intraneural perineurioma.



ARSACS = autosomal recessive spastic ataxia of Charlevoix-Saguenay; CMT1A = Charcot-Marie-Tooth disease type 1A; HNPP = hereditary neuropathy with liability to pressure palsy; MPS = mucopolysaccharidosis.

#### *3.3. Clinical Findings*

Most patients indicated typical complaints of a CTS, e.g., paraesthesia and/or dysaesthesia. However, only 29% (4/14) of CTS patients with MPS indicated complaints related to the CTS, although thenar muscle atrophy was already present at diagnosis in 86% (12/14) of these patients. Thenar muscle atrophy was also present at the time of diagnosis of CTS at a high proportion in most of the other conditions: congenital malformations 3/4, tumors 2/2, traumatic lesion 1/2, rheumatoid arthritis 1/1. Only in CTS associated with neuropathy was thenar muscle atrophy not observed (0/5), and idiopathic CTS showed

thenar muscle atrophy only in 3/9 patients. See Supplementary Table S1 for more detailed individual information.

#### *3.4. Electrophysiological Examination*

Electrodiagnostic studies of all patients revealed a significant latency difference between orthodromic median and ulnar mixed nerve potentials and/or slowing of the antidromic median sensory nerve conduction velocities across the carpal tunnel or lack of the sensory nerve action potential and/or prolonged median DML. See Supplementary Table S1 for detailed information.

#### *3.5. Ultrasound Imaging*

In nine patients US imaging was performed. In all CTS patients, the WFR ratio and/or the CSA of the median nerve at the wrist was increased. In addition, structural lesions of the median nerve were reliably detected. See Supplementary Table S1 for details. In the two patients with perineurioma, the echogenicity and structure of the nerve was altered (enlarged fascicles, increased perineuronal tissue). In these patients an MRI of the wrist and forearm was also performed and confirmed the US findings. The final diagnosis was then confirmed by histological examination following incisional biopsy during surgical decompression of the carpal tunnel.

#### **4. Discussion**

Compared to CTS in adults, CTS in children is rare. However, since children may not present with typical symptoms and may, in part, not communicate their complaints depending on their developmental stage and/or cognitive impairment, CTS is possibly underdiagnosed in this age group. Nevertheless, it is important to consider the presence of CTS even in toddlers with atypical symptoms, because the majority of CTS is caused by an underlying condition and requires early surgical treatment in order to prevent axonal median nerve damage.

In our cohort the age range at diagnosis was 2.5 to 17 years. The youngest child reported with CTS was 9 months old [13]. CTS was bilateral in 76% of our cohort, and a bilateral manifestation occurred mostly in CTS with an underlying hereditary disorder. In comparison, bilateral CTS at manifestation occurs only in approximately 50% of the adult population [14]. A surgical intervention was performed in 71% of our cohort. This high rate of surgical interventions was also related to the secondary nature of childhood CTS. Almost all children from our cohort who harbored a hereditary condition (e.g., lysosomal storage diseases, congenital malformations) or a tumor associated with CTS required surgical intervention because of the low likelihood of improvement under conservative treatment due to the stationary or progressive nature of these conditions.

Mucopolysaccharidosis was the most common cause of CTS in our cohort. A high prevalence of mucopolysaccharidosis in childhood CTS is also reported in the literature [3,8,15]. MPS constitutes a group of rare lysosomal storage diseases with multisystem manifestation. CTS is a common musculoskeletal manifestation of MPS [9,16–18]. The symptoms of CTS in patients with MPS are, however, often not as distinct as in other etiologies. In our cohort, less than 30% indicated complaints related to CTS. The early nonspecific symptoms of CTS in MPS, compounded with communication barriers due to age and intellectual disability, often lead to delayed diagnosis with thenar wasting and potential permanent loss of hand function [17,18]. Therefore, routine biannual physical examination and annual electrophysiological screening for CTS is recommended in the care standards for MPS even in the absence of suggestive symptoms [17]. Adhering to this recommendation, CTS was diagnosed in MPS patients at an early stage in our cohort, and surgical intervention was performed in all MPS CTS patients. Three of 14 MPS patients showed recurrent CTS within 3–11 years after the first surgical intervention. Follow up investigations showed normalization of the nerve conduction studies only in three patients after carpal tunnel release. These three patients were identified and treated very early

(below 5 years of age). The patients with a later surgical intervention showed chronic axonal damage of the median nerve. These findings confirm the importance of physical examinations every six months and annual electrodiagnostic screening, which enables early diagnosis and treatment of CTS in the MPS population. As reviewed by Patel et al. [17], MPS patients are at risk of developing CTS very early in life. In fact, the youngest patient from our cohort, diagnosed with CTS at 30 months of age, belongs to the MPS patient group. Screening for CTS, therefore, should be initiated immediately after the diagnosis of MPS and continued frequently thereafter.

Polyneuropathy was the reason for CTS in five patients (13%) in our cohort. One patient suffered from CMT1A and one patient from sensorimotor neuropathy associated with ARSACS. In one patient, HNPP was assumed as causative for CTS because of a genetically confirmed HNPP in the child's mother. Two additional patients with early onset CTS were siblings, and family history revealed the presence of HNPP over several generations. The parents of these three patients did not consent to the genetic confirmation of HNPP in their children. Del Colle [19] describes an identical constellation compared to the family with two affected siblings in our cohort: A family with HNPP in several generations and a high prevalence of early onset CTS, in some cases as the only manifestation of the HNPP. In general, bilaterally prolonged DML of the median nerve, prolonged DML and/or reduced motor nerve conduction velocities in the peroneal nerve and sensory nerve conduction velocity slowing are highly suggestive of HNPP when there is a positive family history of polyneuropathy [20].

CTS associated with congenital malformations was the fourth most common etiology in our cohort (affecting 11%). Our cohort includes one individual with melorheostosis, an extremely rare and progressive bone disease accompanied by hyperostosis and soft tissue fibrosis. Hand involvement had only been reported sporadically in this condition [21–23]. Interestingly, our cohort also includes two patients with geleophysic dysplasia, a rare hereditary condition characterized by severe short stature, short extremities, progressive joint limitation, thickened skin, and pseudomuscular build. Together with acromicric dysplasia, the geleophysic dysplasia belongs to the acromelic dysplasia group. These two conditions share, in part, similarities of the genetic pathway and phenotype. Hand involvement causes an increased risk for the development of CTS, which might be as high as 35% in geleophysic and acromicric dysplasias [24].

In our cohort we found two patients with an intraneural perineurioma, a rare benign peripheral nerve sheath tumor, which has only been included in the WHO classification system since 2000 [25]. In both patients the diagnosis of the tumor was suspected in the US investigation which followed the electrophysiological diagnosis of CTS. A histopathological examination confirmed the diagnosis in these two patients. Molecular investigation performed with biopsy material showed a pathogenic somatic mutation in the PIK3CA gene in one patient. Perineuriomatous pseudo-onion bulb proliferation is considered part of the PIK3CA-related overgrowth spectrum (PROS) [26] and has also been described in lipomatosis of peripheral nerves with or without nerve territory overgrowth in association with PIK3CA mutations [27]. Dailiana et al. [28] published a case series of tumors and tumor-like lesions affecting the median nerve as rare causes for CTS. However, most of the patients in this study showed nerve compression due to extraneural masses.

In nine patients (24%), the CTS was classified as idiopathic and no obvious underlying condition could be confirmed. This etiological group included one child with a bilateral CTS and the additional diagnosis of familial Mediterranean fever (FMF), who was under colchicine treatment for 18 months prior to the manifestation of bilateral CTS. Since it is known that colchicine can cause polyneuropathies amongst other side effects, Isikay et al. [29] examined a group of 88 children with FMF under Colchicine treatment and found only one patient with CTS. In addition, only Bademci et al. [30] described a bilateral CTS in a young woman with FMF. Due to the fact of the high incidence of both FMF in some populations and CTS in general, this association might be random, which is also the conclusion of a large retrospective study of comorbidities in 2000 FMF patients including

more than 600 children [31]. Our patient with bilateral CTS and FMF might nevertheless be an example for the suspicion that even the low proportion of idiopathic CTS in children might be overestimated, because this etiological group most likely includes patients with underlying disorders that are unknown or not detectable at the time of CTS manifestation.

Neuromuscular US is becoming a standard investigation in the evaluation of peripheral nerve and muscle diseases, including CTS. In addition to electrodiagnostic procedures, detection of median nerve enlargement at the wrist by US has been suggested as a sensitive and valuable diagnostic method [32–34]. Billakota et al. [6] performed a large retrospective analysis of median nerve ultrasound investigations in CTS and concluded that median nerve US is nearly as sensitive as electrophysiological testing, which is considered the diagnostic gold standard in CTS. Bäumer et al. [35] specifically examined the value of US in the management of patients with MPS. In their study, US had an even higher sensitivity for the detection of CTS compared to electrophysiology. In our cohort, which focused on the etiology and not on the diagnostic measures, US was performed only in a small number of patients. This is primarily a consequence of the retrospective nature of this study. Even the low number of median nerve US investigations in our cohort, however, demonstrates that the increase of both the CSA of the median nerve at the wrist and the WFR are also indicative of CTS in children. In addition, nerve US is a sensitive method to detect tumor associated median nerve lesions. Therefore, US is a valuable tool to support the clinical and electrophysiological diagnosis of CTS, especially in children since it is a quick, non-invasive and painless method.

In conclusion, we were able to identify a broad spectrum of underlying etiologies in our cohort of childhood CTS. Our study confirms that idiopathic CTS in children is rare and most commonly secondary to an underlying condition with mucopolysaccharidosis as the most common cause.

To prevent delayed diagnosis, which can lead to thenar wasting and permanent loss of hand function, we propose the following diagnostic algorithm for patients at risk and patients with symptoms suggestive of CTS:

Patients with conditions associated with a high risk for CTS should be clinically screened for symptoms of CTS at the time of the primary diagnosis and frequently thereafter. Thenar muscle atrophy, pain and/or sensory symptoms, and/or disturbances of nail growth in digits I to III, a positive Tinel sign at the wrist and/or deterioration of dexterity are features potentially pointing to the presence of CTS. Physical examination is recommended as frequent as every 6 months for MPS patients, the largest patient group at risk for CTS [17,18]. In addition to lysosomal storage diseases, several genetic conditions, including HNPP and other hereditary neuropathies and congenital malformation syndromes like acromelic dysplasia, melorheostasis, and hemihypertrophia syndromes have a high risk for early CTS, too, and should also be screened for CTS by frequent physical examination. It is, however, important to emphasize, that CTS symptoms in children are often atypical and complaints might not be communicated due to the developmental stage and/or cognitive impairment of these patients. Therefore, physical examination and screening for CTS should be supplemented by median nerve US. Our study shows that the early conduct of median nerve US might be diagnostic for childhood CTS. In addition, all patients at risk should undergo annual or even more frequent electrodiagnostic investigations, applying standard procedures for the investigation of CTS. If these investigations prove the presence of CTS, early surgical intervention should be discussed because conservative treatment might not be effective due to the stationary or progressive nature of the primary conditions and the high risk of axonal median nerve damage.

Patients with symptoms suggestive of CTS without known underlying conditions should undergo the same procedures consisting of physical examination, median nerve US, and electrodiagnostic testing. Due to the frequent secondary nature of childhood CTS, a thorough investigation and search for an underlying disease is mandatory.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/children8080624/s1, Table S1: Demographics, etiology, and clinical and diagnostic findings of childhood CTS.

**Author Contributions:** Conceptualization, C.T.R., U.K., P.J.B. and G.M.S.; data collection and analysis, C.T.R., U.K., M.R. and G.M.S.; clinical investigations, all authors; electrodiagnostic investigations, U.K., J.L., P.J.B. and G.M.S.; US investigations, A.E. and P.J.B.; surgical treatment D.M.W. and K.W.; writing—original draft preparation, C.T.R., U.K. and G.M.S.; writing—review and editing, all authors; funding acquisition, G.M.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by a training fellowship granted by the Swiss Muscle Society to CT Rüsch.

**Institutional Review Board Statement:** The study was approved by the local ethics committee on 23 June 2020, and registered with the Swiss project database (BASEC 2020-01016).

**Informed Consent Statement:** Written informed consent was obtained from the caregivers of all subjects prior to inclusion of participants in the study.

**Data Availability Statement:** Data is contained within the article or Supplementary Material.

**Acknowledgments:** We thank all patients and caregivers for their participation in this study.

**Conflicts of Interest:** The authors declare no conflict of interest. The sponsor had no role in the design, execution, interpretation, or writing of the study.

#### **References**

