Nerve ultrasound characterizes AMN polyneuropathy as inhomogeneous and focal hypertrophic

High-resolution ultrasound (HRUS) is an inexpensive, quick, and comfortable tool to screen for nerve cross-sectional area (CSA) changes. These changes range from focal to multifocal, or homogenous enlargements [1] even without electrophysiological changes. Despite obvious advantages of HRUS to identify and evaluate structural peripheral nerve alterations, no data for X-linked adrenoleukodystrophy (X-ALD) exists. X-ALD is the most common peroxisomal disorder worldwide [2]. The disease is caused by impaired peroxisomal beta-oxidation due to mutations in the ABCD1 gene [3] on the X-chromosome leading to an accumulation of very-long chain fatty acids (VLCFA) in plasma as well as tissues including white matter of the brain, spinal cord, and adrenal cortex. Patients with adult onset usually develop the adrenomyeloneuropathy (AMN) phenotype characterized by progressive spastic paraparesis with bladder disturbance, sensory ataxia with impaired vibration sense, pain in the legs and in male patients, adrenal failure. Nerve conduction studies (NCS) mostly reveal multifocal demyelination [4] with lower extremities most frequently and most severely affected [5]. Central findings [6] include prolonged central somatosensory conduction and prolonged central motor conduction.

On examination, clinical signs of potential peripheral neuropathy were common. 92% of patients had afferent ataxia, 85% showed distally pronounced weakness of the lower limbs, and 62% had malleolar pallhypaesthesia. Neither nerve conduction studies (NCS) nor ultrasound findings correlated with clinical signs of neuropathy, with disease duration, or disease severity as assessed by the SPRS [13]. Detailed clinical as well as genetic data of all patients is presented in Additional file 1: Table S1.

Nerve ultrasound found CSA of peripheral nerves to be enlarged in 6/13 patients. Nerve enlargement was mostly inhomogeneous or regional (Fig. 1 (Ia&b) and Table 1). CSA enlargement was most prominent in the proximal upper extremities (median and ulnar nerve) and the vagal nerve. It was also found in the C5 root, the sural nerve, and the distal segment of the tibial nerve (Table 1). The proximal to distal ratio showed significant differences for the median nerve (t-test with Bonferroni p = 0.021 with larger ratios in AMN with neuropathy (median ratio 2.0, range 1.5–3.6) compared to patients without neuropathy (median ratio 1.3, range 0.9–1.5) and controls (median ratio 1.3, range 1.0–1.7)), with no significant findings in the ulnar and tibial nerve. NCS revealed significantly reduced motor nerve conduction velocity (CV) in 5/13 patients with AMN ranging from 23 to 42 m/s in the UN and from 20 to 36 m/s in the TN while in eight AMN patients NCS were unremarkable (Additional file 1: Table S1). Interestingly, nerve ultrasound findings were abnormal in all AMN patients with electrophysiological indicators of demyelinating peripheral neuropathy. Vice versa, ultrasound measurements were normal (Fig. 1 – pictures IIa&b) in all but one AMN patient (#3) without neuropathy in NCS; in this patient (#3) CSA enlargement was restricted to the lower limbs (Additional file 1: Table S1). CSA showed close inverse correlation with motor nerve CV of the tibial and ulnar nerve in all patients (Fig. 2a). CSA did not correlate significantly with sensory nerve CVs of the ulnar (p = 0.142) and sural nerve (p = 0.173). The ultrasound pattern sum score (UPSS) as overall nerve enlargement score, its subscores and the HS revealed significant differences between AMN patients with and without neuropathy.

By ROC-curve analysis a cut-off value > 3 in the UPSS was evaluated to have the best sensitivity and specificity to differentiate between AMN patients with and without electrophysiological evidence of polyneuropathy (Fig. 2b).

The HS showed significantly larger results in AMN with neuropathy compared to those without (p = 0.004) and the healthy controls (p = 0.002) tested by t-test with Bonferroni correction. Although median values of CSA and UPSS were slightly higher in males than in females, there was no significant gender difference. Similarly, no effect of age, height, or weight on ultrasound data was detected (all parameters tested with Mann-Whitney-test). All patients with AMN and nerve enlargement showed reduced echointensity of the fascicles corresponding to Class 1 pattern described by Padua and colleagues [11] (Fig. 1, Ia and b). All but one patient with AMN and no electrophysiological neuropathy revealed Class 3 pattern.

In our AMN cohort significant electrophysiological neuropathy was found in about 50% of cases equally affecting males and females. This is less than the clinical findings suggested but central affection can lead to an overestimate of peripheral nerve involvement. If present, AMN-associated neuropathy shows demyelinating characteristics in NCS as previously described [4]. HRUS revealed AMN neuropathy being accompanied by regional and predominantly proximal hypertrophy, particularly of the median nerve and radices. We propose that increased nerve CSA represents the primary morphological equivalent of AMN neuropathy. When comparing HRUS findings in AMN to other demyelinating polyneuropathies like Charcot-Marie-Tooth type 1a (CMT1a) or chronic inflammatory demyelinating polyneuropathy (CIDP), striking differences stand out (see Fig. 1): CMT1a typically leads to homogeneous nerve enlargement in HRUS, homogeneous decreased CV in motor and sensory NCS, and onion-bulb formations in histopathology caused by chronic de- and re-myelination. Both HRUS scores, UPSS and the homogeneity score (HS) are obviously higher in CMT1a (as described in [10]) than in AMN. In CMT1a the echointensity is reduced with swollen fascicles to the best of our knowledge and thus would be classified as Class 1 according to Padua et al.; however, we must admit, that this classification has never been used for hereditary neuropathies so far. In CIDP, HRUS depicts rather inhomogeneous nerve enlargement with predominance in roots and nerve sections at the upper arm. In contrast to AMN (Padua Class 1 pattern according to [11]), increased nerve echo as a potential ultrasound sign of intraneural fibrosis might be a typical finding in many CIDP patients, particularly in those with long disease duration [11, 15]. Härtig and colleagues correlated histology of those patients with increased echointensity (Padua Class 2 pattern) and demonstrated more axonal damage in those nerve biopsies. NCS typically finds regionally restricted slowing of CV with prolonged F-wave latency, temporal dispersion or conduction blocks in CIDP. Histopathology reveals onion-bulb formations [2] in combination with inflammation, edema, fibrosis and axonal damage [16].

Taken together, differences can be summarized as following (compare Fig. 1):

Due to the limited number of nerve biopsies and the restriction to the distal sural nerve of patients with AMN it is difficult to define the histopathological basis of HRUS abnormalities in AMN. In contrast, skin biopsy data showing lack of small nerve fibers in patients with AMN is available [17]. The predominant involvement of proximal nerve segments of the upper extremities (particularly the median nerve) is noteworthy, as similar findings have been described for acquired inflammatory neuropathies, i.e. CIDP [18] or MMN [19] and even inherited neuropathies, i.e. Friedreich’s ataxia [20] or familial amyloidosis [21]. The pathophysiological background of this finding has not been clarified until now, but proximal nerve segments might be more vulnerable than distal ones. Thickening of nerves might be a sign of concomitant inflammation – as in CIDP – which could even pinpoint to therapeutic steps, however this needs to be clarified by additional investigations, i.e. CSF analysis, MRI with gadolinium, biopsies or post mortem tissue. Another explanation could be a type of hypertrophic remodeling, which has already been discussed before for inflammatory and inherited neuropathies [22].

Predisposing factors of neuropathy in AMN are not known [2] and the evolution of AMN neuropathy is still enigmatic. There were no correlations with gender, age, height, weight, or disease severity. It is unclear to which extent the demyelinating process found in our case series results in axonal damage during further disease progression. This is well established for other hereditary demyelinating neuropathies like CMT1a. Furthermore, it is unclear whether structural nerve changes as recognized in HRUS precede electrophysiological changes or if they occur later (as it is potentially the case in patient #3). Longitudinal follow-up with repeated ultrasound examination can help to answer these questions. Interestingly, nerve swellings are more prominent in the upper extremities. As ultrasound changes do not correlate with disease duration, age, or disease severity there is no hint that this reflects secondary atrophy of distal parts of the nerves after initial swelling. Better understanding of the development of AMN neuropathy is needed and HRUS may aid or guide histopathological understanding to further advances.

Nerve ultrasound is a reliable, painless, and easily accessible tool to characterize peripheral nerve involvement in AMN patients. Our findings have profound practical implications because neuropathy in AMN patients is characterized by multifocal proximal nerve enlargements with reduced echo intensity in high resolution ultrasound and homogeneous reduction of conduction velocity in nerve conduction studies. Ultrasound and NCS characteristics in AMN seem to differ from other demyelinating neuropathies. As ultrasound is a fast examination technique that does not cause discomfort to the patients it has compliance advantages compared to electrophysiological examinations. Long-term follow-up with HRUS will help to disclose the evolution of peripheral neuropathy in patients with AMN.