Monomelic amyotrophy of the upper limb or Hirayama disease is an anterior horn disorder that is restricted to one upper extremity in the majority but can be bilaterally symmetrical in a minority of cases.1,2 The classical features include an insidious arm weakness and atrophy starting in the second or third decade of life, a male predominance, a sporadic occurrence without family history, and a slow progression followed by stabilization resulting in a benign outcome. The electromyographic findings are suggestive of active denervation and chronic reinnervation in the affected arm and at times the clinically uninvolved contralateral arm. On magnetic resonance imaging (MRI), there is often a forward compression of lower cervical spinal cord against the anterior spinal canal wall in association with an anterior shift of the posterior dural sac during neck flexion.2 It is believed that such an anterior horn disorder results from local arterial compression and ischemia in the anterior section of spinal cord. However, the imaging finding of dural shifting is nonuniversal, and another possible mechanism of intrinsic motor neuron degeneration has often been suggested.3–5 Here we describe a longstanding case of Hirayama disease whose cervical spine MRI demonstrated a “snake-eye” appearance on axial images but without dura shifting. We reviewed the literature and compared the image findings with a variety of cervical myelopathies, secondary to vascular insufficiency, including vertebral dissection, arterial embolism, atherosclerotic disease, and hypoperfusion. We believe that the disorder can still be explained by repetitive arterial compression leading to anterior spinal cord ischemia in patients with Hirayama disease even without evidences of dural shifting on MRI.
At 14 years of age, a white boy developed slowly progressive difficulty in using his left hand without pain or numbness. At 19 years, examination showed a mild to moderate weakness in the left finger extensors, abductor pollicis longus, abductor digiti minimi, and first dorsal interosseous muscles. Mild atrophy was noticed in the left first dorsal interosseous region. Strength in the other muscle groups of the left arm and the remaining 3 limbs were completely normal. All tendon reflexes and sensory examination were normal. Investigations including creatine kinase, sedimentation rate, cryoglobulin, and serum protein electrophoresis were normal. Cerebrospinal fluid study and cervical computed tomography myelogram were both normal. Nerve conduction studies in the upper extremities were normal except for a left ulnar compound muscle action potential of slightly reduced amplitude. Needle examination revealed the presence of spontaneous activities and motor unit potentials of large amplitude, long duration with reduced recruitment in the left first dorsal interosseous, abductor pollicis brevis, and extensor digitorum communis muscles. Similar but more subtle changes were seen in the right abductor pollicis brevis and first dorsal interosseous muscles. Over the next 26 years he felt only a slight worsening of the strength and coordination of his left hand with minimal functional limitation.
At 45 years, he demonstrated a normal mental status and cranial nerve examination. Muscle strength in the right arm and bilateral lower extremities remained normal. Muscle strength in the left upper extremity was recorded as follows (Medical Research Council scale 0–5): first dorsal interosseous, 3; abductor pollicis brevis, 2; extensor digitorum communis, 4; and flexor pollicis longus, 4. Strength examinations in the left brachioradialis, deltoid, biceps, and triceps were all normal. Significant atrophy was noticed in the left thenar, hypothenar, and interosseous regions. Reflexes were normal in the biceps but slightly hyperactive in the bilateral triceps, knee, and ankle tendon reflexes. No pathological reflexes were recorded, and sensory examination was completely normal. There has never been a history of major trauma. On repeated nerve conduction studies, the left median and ulnar compound muscle action potentials showed significantly reduced amplitudes and slightly reduced conduction velocities. The needle examination findings showed more significant neurogenic changes of motor unit potentials in the same muscle groups as 26 years ago.
Other investigations included a computed tomography angiogram showing normal bilateral vertebral arteries, and a normal bilateral upper extremity arterial Doppler study. MRI of the cervical spine showed segmental cord atrophy at C5 through C7 levels. Instead of being a round configuration, the anterior border of the spinal cord demonstrated a concave deformity at these levels, suggesting anterior compression. On axial image, intrinsic cord signals were seen in the anterior segment, more pronounced on the left than right, with a typical snake-eye appearance. On sagittal images, these signal abnormalities occurred at the C4-5, C5-6, and C6-7 levels corresponding precisely to levels of the intervertebral disks, whereas cord segments behind vertebral bodies showed minimal or no signal abnormalities. No central canal stenosis or cord impingement was seen at the C4-5, C5-6, and C6-7 levels at the neutral position. Sagittal images upon neck flexion did not reveal an anterior shifting of the posterior dura or the appearance of a prominent epidural contrast enhancement. However, narrowed ventral subarachnoid space was seen at the C5-6 and C6-7 levels upon flexion, more pronounced on the left than right side (Fig. 1). No significant changes were noticed on MRI over a period of 7 years.
The patient's clinical course and electrophysiological findings fit the diagnosis of monomelic amyotrophy of the upper limb or Hirayama disease. The pathogenesis of Hirayama disease includes a forward compression of the lower cervical spinal cord occurring against the anterior wall of the spinal canal because of an anterior shift of the posterior dural sac upon neck flexion. In flexion, the dural canal becomes tightened and displaced anteriorly because of a disproportional lengthening of the vertebral canal and a relative shortening of the posterior dural sac.6,7 The anterior shifting and flattening of the spinal cord were also observed during intraoperative ultrasonography.8 Repeated or sustained neck flexion renders the anterior spinal cord into chronic circulatory insufficiency. Using spinal angiography to study 2 patients, Elsheikh et al9 did not detect blood flow impairment in the major arteries upon flexion. This suggests that large vessel obstruction does not occur in Hirayama disease, and circulatory impairment may occur at the arterial branches or microcirculatory levels. Classical MRI findings in Hirayama disease include the following: (1) focal atrophy in the lower cervical cord; (2) anterior shifting of the posterior dura; (3) venous plexus dilatation appearing as a crescent or focal mass with flow voids in the enlarged posterior epidural space displaying strong contrast enhancement, perhaps as a result of negative pressure created by the forceful movement of posterior dural sac (Fig. 2).10,11 The nature of these epidural masses as engorged venous plexuses has been confirmed on angiography.9,12 The compression of cervical spinal cord is often asymmetric as an explanation for the predominantly unilateral presentation. The male preponderance is perhaps because of the shorter length of cervical spinal canal and spinal cord seen in women.8 The self-limiting course is probably explained by the less arterial compression after the development of local cord atrophy, and there has been an observation of less or no dural shifting in patients with a disease course of more than 10 years.10 In large study series, 87% and 95% patients demonstrated a forward displacement and flattening of lower cervical cord and a posterior venous engorgement on MRI.10,13 Histological studies suggested the posterior dura in these patients possesses a reduced content of elastin and becomes inelastic, which could be a predisposition factor in the development of Hirayama disease.8
Contrary to the above observations, several authors reported a lack of dural shifting and the presence of normal posterior epidural spaces upon flexion in a minority of patients. Based on these observations, these authors proposed that Hirayama disease is perhaps a degenerative motor neuron disease rather than a flexion myelopathy.3–5 Schroder et al3 measured the anteroposterior spinal cord diameter in each vertebral segment and did not find significant difference in the extent of spinal cord flattening between patients and healthy control subjects. However, the average duration of illness in their series was more than 14 years. After such a chronic course, the spinal cord atrophy become permanent and significant morphological change may no longer be seen in association with dynamic changes. Willeit et al4 reported 3 cases demonstrating high-intensity signals in the anterior horn cell region of the lower cervical spinal cord on MRI but without dural shifting, similar to our case. Ammendola et al5 described 3 cases, but only one of the 3 underwent dynamic MRI imaging. The imaging of this patient failed to show an anterior shifting of the posterior dural sac upon flexion. However, significant narrowing of anterior subarachnoid space, widening of posterior subarachnoid space, and direct contact of spinal cord with anterior spinal wall were evident at the lower cervical spine upon flexion in the figures of both publications.4,5 In a process of a dynamic compression, the position of cervical cord and the degree of cord compression upon flexion may also vary depending on whether patients are prone or supine. Most MRI studies are performed while patients are lying supine. More significant anterior cord compression was seen when patients were placed in a prone position because of the gravity factor.14
Other lines of evidences exist in supporting a mechanism of compression and ischemia. Laboratory findings suggested that anterior horn is most vulnerable to chronic ischemia, whereas spinal cord white matter is more resistant. Autopsies performed on 2 patients with Hirayama disease supported circulatory insufficiency in the territory of anterior spinal cord.15,16 Decreased spinal cord pulsation was noted on intraoperative ultrasonography upon neck flexion.8 Reports on the treatment of Hirayama disease suggested a slowing down of progression or even improvement after cervical collar therapy or surgical decompression.8,16,17 Limited molecular analysis on Hirayama disease thus far does not support a genetic contribution. Deletion analysis of the survival motor neuron (SMN) genes was normal in a study of 15 patients.18 Sequences of the superoxide dismutase 1 (SOD1) gene in 2 patients were normal as well.19
The cervical spine MRI of this patient did not reveal a dural shifting upon neck flexion. However, typical findings of Hirayama disease such as significant focal atrophy were noted in the lower anterior cervical spinal cord, consistent with many previous reports (Fig. 1).10 In addition, intrinsic signal changes occurred at multiple intervertebral disk levels rather than being adjacent to the posterior border of vertebral bodies. No significant central canal stenosis or cord impingement was noticed at neutral position, but an anterior compression of the spinal cord was detected at levels of protruding disks upon flexion. The intensity of signal changes seemed to correspond to the severity of disk compression (Fig. 1). Similar locations of signal changes corresponding to disk levels have been shown previously in patients with Hirayama disease.20 On the axial images, there was a snake-eye appearance that was previously reported in Hirayama diseases with and without evidences of dural shifting.2,4 Such an MRI appearance was also described for acute spinal cord ischemia, such as hypoperfusion, aortic diseases including atherosclerosis, aneurysm, thrombosis, dissection or surgical repair, and dissection of vertebral arteries.21–24 Therefore, we feel that the best explanation for the pathogenesis of our patient is a state of chronic recurrent ischemia in the anterior spinal cord segment associated with repetitive neck flexion because of a compression of the anterior spinal or radicular arterial branches by protruding disks rather than an anterior shifting of posterior dura. The observed disk protrusions in this patient appeared slight and nonharmful at the neutral position. However, more significant anterior compression was seen upon neck flexion while lying supine that could possibly become even more obvious in a prone position. Future MRI studies in Hirayama disease may need to include examining patients on a prone rather than supine position, especially for patients without characteristic dural shifting.
In summary, Hirayama disease is a clinical syndrome of cervical flexion myelopathy because of the compression of spinal cord against the anterior spinal canal including vertebral bodies and intervertebral disks, with or without an anterior shift of the posterior dural sac. However, Hirayama disease should be differentiated from the traditional cervical spondylotic myelopathy. The latter is often seen in patients of advanced age and significant central canal stenosis is often observed at neutral position. There tends to be a course of acute or fast progression. Physical examination often reveals sensory disturbance or long tract signs. Although surgery is indicated and beneficial for cervical spondylotic myelopathy, the role of dural decompression or duraplasty in Hirayama disease warrants further large-scale studies.
1. Hirayama K. Juvenile muscular atrophy of distal upper extremity (Hirayama disease
). Intern Med. 2000;39:283–290.
2. Pradhan S. Bilaterally symmetric form of Hirayama disease
. Neurology. 2009;72:2083–2089.
3. Schröder R, Keller E, Flacke S, et al.. MRI findings in Hirayama's disease: flexion-induced cervical myelopathy or intrinsic motor neuron disease?. J Neurol. 1999;246:1069–1074.
4. Willeit J, Kiechl S, Kiechl-Kohlendorfer U, et al.. Juvenile asymmetric segmental spinal muscular atrophy (Hirayama's disease): three cases without evidence of “flexion myelopathy”. Acta Neurol Scand. 2001;104:320–322.
5. Ammendola A, Gallo A, Iannaccone T, et al.. Hirayama disease
: three cases assessed by F wave, somatosensory and motor evoked potentials and magnetic resonance imaging not supporting flexion myelopathy. Neurol Sci. 2008;29:303–311.
6. Kikuchi S, Tashiro K, Kitagawa M, et al.. A mechanism of juvenile muscular atrophy localized in the hand and forearm (Hirayama's disease)—flexion myelopathy with tight dural canal in flexion. Rinsho Shinkeigaku. 1987;27:412–419.
7. Kohno M, Takahashi H, Yagishita A, et al.. “Disproportion theory” of the cervical spine and spinal cord in patients with juvenile cervical flexion myelopathy. A study comparing cervical magnetic resonance images with those of normal controls. Surg Neurol. 1998;50:421–430.
8. Konno S, Goto S, Murakami M, et al.. Juvenile amyotrophy of the distal upper extremity: pathologic findings of the dura mater and surgical management. Spine (Phila Pa 1976). 1997;22:486–492.
9. Elsheikh B, Kissel JT, Christoforidis G, et al.. Spinal angiography and epidural venography in juvenile muscular atrophy of the distal arm “Hirayama disease
”. Muscle Nerve. 2009;40:206–212.
10. Ciceri EF, Chiapparini L, Erbetta A, et al.. Angiographically proven cervical venous engorgement: a possible concurrent cause in the pathophysiology of Hirayama's myelopathy. Neurol Sci. 2010;31:845–848.
11. Chen CJ, Chen CM, Wu CL, et al.. Hirayama disease
: MR diagnosis. AJNR Am J Neuroradiol. 1998;19:365–368.
12. Hirayama K, Tokumaru Y. Cervical dural sac and spinal cord in juvenile muscular atrophy of distal upper extremity. Neurology. 2000;54:1922–1926.
13. Huang YC, Ro LS, Chang HS, et al.. A clinical study of Hirayama disease
in Taiwan. Muscle Nerve. 2008;37:576–582.
14. Breig A, Turnbull I, Hassler O. Effects of mechanical stresses on the spinal cord in cervical spondylosis. A study on fresh cadaver material. J Neurosurg. 1966;25:45–56.
15. Araki K, Ueda Y, Michinaka C, et al.. An autopsy case of juvenile muscular atrophy of unilateral upper extremity (Hirayama's disease). Nihon Naika Gakkai Zasshi. 1989;78:674–675.
16. Hirayama K. Juvenile muscular atrophy of distal upper extremity (Hirayama disease
): focal cervical ischemic poliomyelopathy. Neuropathology. 2000;20(suppl):S91–S94.
17. Lin MS, Kung WM, Chiu WT, et al.. Hirayama disease
. J Neurosurg Spine. 2010;12:629–634.
18. Misra UK, Kalita J, Mishra VN, et al.. A clinical, magnetic resonance imaging, and survival motor neuron gene deletion study of Hirayama disease
. Arch Neurol. 2005;62:120–123.
19. Robberecht W, Aguirre T, Van den Bosch L, et al.. Familial juvenile focal amyotrophy of the upper extremity (Hirayama disease
). Superoxide dismutase 1 genotype and activity. Arch Neurol. 1997;54:46–50.
20. Pradhan S, Gupta RK. Magnetic resonance imaging in juvenile asymmetric segmental spinal muscular atrophy. J Neurol Sci. 1997;146:133–138.
21. Mawad ME, Rivera V, Crawford S, et al.. Spinal cord ischemia
after resection of thoracoabdominal aortic aneurysms: MR findings in 24 patients. AJR Am J Roentgenol. 1990;155:1303–1307.
22. Hundsberger T, Thömke F, Hopf HC, et al.. Symmetrical infarction of the cervical spinal cord due to spontaneous bilateral vertebral artery dissection. Stroke. 1998;29:1742.
23. Weidauer S, Nichtweiss M, Lanfermann H, et al.. Spinal cord infarction: MR imaging and clinical features in 16 cases. Neuroradiology. 2002;44:851–857.
24. Masson C, Pruvo JP, Meder JF, et al.; Study Group on Spinal Cord Infarction of the French Neurovascular Society. Spinal cord infarction: clinical and magnetic resonance imaging findings and short term outcome. J Neurol Neurosurg Psychiatry. 2004;75:1431–1435.