Keywords Hirayama disease - paraspinal muscles - morphometric - flexor–extensor disparity
Key Message
Though the clinical manifestations of Hirayama disease are limited to the lower cervical
segments, this study demonstrates morphometric alterations at all levels of the subaxial
cervical paraspinal musculature. The flexor–extensor muscle disparity in these patients
could be utilized as a potentially modifiable factor in the conservative management
protocol of the disease.
Introduction
Monomelic amyotrophy, commonly known as Hirayama disease (HD) or juvenile muscular
atrophy of the distal upper extremities, is an unusual disorder characterized by self-limiting
weakness and atrophy of the hands and forearm muscles in young male subjects.[1 ]
[2 ]
[3 ] Flexion-induced myelopathy is one of the widely accepted pathogenetic mechanisms,[1 ]
[4 ]
[5 ] though the exact pathogenesis of the disease remains controversial. The clinical
manifestations of HD are limited to the lower cervical segments. However, recent studies
have demonstrated biomechanical connotations of the disorder that relate to the entire
cervical spine. Given this background and the fact that paraspinal musculature is
the major biomechanical stabilizer of the cervical spine,[6 ] we postulated that patients with HD would have morphological alterations in their
paraspinal muscles (PSMs). This has been previously demonstrated, though only at select
lower cervical levels.[7 ] The objective of the present study was to explore if morphometric changes in HD
extend to other levels of the subaxial cervical spine as well. We felt that a comprehensive
evaluation of the entire subaxial cervical musculature was especially required for
a condition where long-term-collar wear is a primary management protocol despite its
potentially deleterious effects on PSMs and hence, on cervical biomechanics.
Subjects and Methods
Patient Population
This was a single-center case–control study spanning an 8-year-period from January
2012 to January 2020. In view of the retrospective and anonymized nature of the data,
the study qualified for an ethics committee waiver as per the local Institutional
Review Board policy. The study group consisted of 64 consecutive patients diagnosed
to have HD at our institution based on the criteria proposed by Hirayama et al[8 ]
[9 ]: weakness and wasting in the C7-T1 myotomes in one or both upper limbs, insidious
onset of symptoms in the early third decade, rapid progression of symptoms from 1
to 3 years, irregular coarse tremors in the fingers of the affected hand(s), brief
aggravation of symptoms on exposure to the cold, electromyographic evidence of chronic
denervation in the affected muscles, and an absence of objective sensory loss. None
of these patients were previously managed with collar wear for any cervical pathology,
nor did they have history of prior trauma or cervical surgery.
Control Group
Using an individual, “one-to-one” matching technique,[10 ] 64 age- and sex-matched non-HD subjects were selected as controls for comparison
of their PSM morphometry with that of the study group. The controls had undergone
cervical spine imaging as part of a routine neurological evaluation, and none of them
had history of neck-pain, radicular pain, cervical trauma, or previous neck surgery.
Radiographic Evaluation
The same imaging protocol was used for both, the study and control groups. Deidentified
magnetic resonance (MR) images were obtained from the hospital radiographic system
(Synapse, Fujifilm Medical Systems USA, Inc., Lexington, Massachusetts, United States).
All images were acquired on a HDi 1.5 Tesla magnet (GE Signa, Milwaukee, Wisconsin,
United States) using a standard neurovascular (NV) coil. Measurement parameters were
slices, 26; slice thickness, 3.7 mm; field of view, 180 × 188 mm; repetition time
(TR), 4,360 milliseconds; time to echo (TE), 98.2 milliseconds; matrix size, 1.60/256;
number of excitations, 1.5; flip angle, 90 degrees. The cross-sectional areas (CSAs)
of the PSMs were measured by a standardized technique using axial T2-weighted images.[11 ]
[12 ] The selected cuts were parallel to the disc spaces at the upper endplates from C3
to C7 vertebral levels. A region of interest (ROI) was created for each muscle bilaterally
using the Synapse software. The lateral border of the facets was used as the lateral
limit for the ROI for the superficial extensors (SEs). The CSAs of the following muscles
were measured: sternocleidomastoid (superficial flexor, SF), longus colli and longus
capitis (deep flexors, DFs), multifidus and semispinalis cervicis (deep extensors,
DEs), and the semispinalis capitis, splenius capitis, and upper trapezius (SE) ([Fig. 1 ]). Two independent operators (PA and NR) performed all the measurements, and the
mean values of their measurements were used for analysis. Both the operators were
blinded to the group allocation into cases and controls. To control for bias in the
muscle areas due to varying body mass index, muscle CSA/vertebral body area (VBA)
ratios were used rather than absolute muscle CSA values.[12 ] The mean CSA/VBA ratio at individual levels and for the entire subaxial spine were
calculated for each muscle group.
Fig. 1 T2-weighted axial magnetic resonance imaging section demonstrating the measurement
of the cross-sectional area of the paraspinal muscles, with regions of interest drawn
for the superficial and deep muscle groups. DE, deep extensor; DF, deep flexor; SE,
superficial extensor; SF, superficial flexor; VBA, vertebral body area.
Statistical Methodology
Using the clinically meaningful difference in CSA values from a previous study on
cervical PSMs,[12 ] a sample size of 64 in each group was determined to be adequate to obtain a power
of 0.9 for the study. Data was entered in an Excel spreadsheet (Microsoft Inc.) and
analyzed using SPSS version 20. Means and standard deviations were computed for all
the CSA/VBA ratios. Student's t -test was used to compare differences between the ratios in the study and control
groups. Inter-observer variability was measured using the intraclass correlation coefficient
and standardized ratings of agreement.
Results
Demographics and clinicoradiological presentation of the study group : All the patients were male subjects with a mean age of 20.49 ± 2.87 years. Their
mean duration of symptoms was 24.36 ± 8.32 months (range: 14–48 months). Forty-eight
(75%) patients presented with bilateral, asymmetric weakness, or wasting of the hand
and/or forearm muscles, while the remaining 16 (25%) presented with unilateral symptoms
and signs. All patients demonstrated lower cervical cord atrophy on neutral MRI and
forward migration of the cord with a prominent enhancing epidural venous plexus on
flexion sequences ([Fig. 2 ]).
Fig. 2 Sagittal magnetic resonance imaging: (A ) T2-weighted sequence demonstrating lower cervical cord atrophy; (B ) Gadolinium-enhanced flexion sequence demonstrating the characteristic enhancing
venous plexus.
CSA/VBA Ratios at Individual Levels
The mean CSA/VBA ratios at individual levels from C3 to C7 for the respective muscle
groups are listed in [Table 1 ]. Patients with HD uniformly demonstrated larger mean SF and DF CSA/VBA ratios than
the controls at all levels, with the differences being significant at most levels.
The extensor muscle CSA/VBA ratios, on the other hand, were smaller in HD patients
than in the controls. The difference in the CSA/VBA values for the DEs in the two
groups was significant at all levels.
Table 1
Paraspinal muscle CSAs at different cervical levels compared between cases and controls
Level
Muscle group
Mean CSA/VBA ratio
p -Value
Cases (n = 64)
Controls (n = 64)
C3
SF
1.14 ± 0.16
0.99 ± 0.15
< 0.0001
DF
0.56 ± 0.08
0.45 ± 0.08
< 0.0001
SE
0.47 ± 0.23
0.50 ± 0.21
0.44
DE
1.14 ± 0.20
1.26 ± 0.02
< 0.0001
C4
SF
1.32 ± 0.34
1.25 ± 0.45
0.32
DF
0.55 ± 0.10
0.47 ± 0.13
0.0002
SE
0.63 ± 0.07
0.65 ± 0.50
0.75
DE
1.09 ± 0.07
1.30 ± 0.07
< 0.0001
C5
SF
1.42 ± 0.01
1.33 ± 0.20
0.0005
DF
0.45 ± 0.22
0.44 ± 0.03
0.71
SE
0.76 ± 0.01
0.82 ± 0.19
0.01
DE
0.97 ± 0.15
1.19 ± 0.28
< 0.0001
C6
SF
1.33 ± 0.22
1.30 ± 0.18
0.40
DF
0.39 ± 0.19
0.34 ± 0.02
0.03
SE
0.75 ± 0.14
0.82 ± 0.23
0.03
DE
0.77 ± 0.05
0.92 ± 0.17
< 0.0001
C7
SF
1.28 ± 0.30
1.16 ± 0.19
0.007
DF
0.61 ± 0.46
0.54 ± 0.29
0.30
SE
0.62 ± 0.26
0.86 ± 0.01
< 0.0001
DE
0.98 ± 0.11
1.06 ± 0.01
< 0.0001
Abbreviations: CSA, cross-sectional area; DE, deep extensors; DF, deep flexors; SE,
superficial extensors; SF, superficial flexors; VBA, vertebral body area.
Overall subaxial spine CSA/VBA ratios : Patients with HD demonstrated significantly larger mean SF and DF CSA/VBA ratios
for the entire subaxial spine ([Fig. 3 ]), while their mean SE and DF ratios were significantly smaller than those of the
control subjects. The overall SF/SE, DF/DE, and total flexor/total extensor ratios
were correspondingly larger in the HD patients than in the controls ([Fig. 4 ]).
Fig. 3 Error-bar chart demonstrating the comparison of the various paraspinal muscle cross-sectional
area/vertebral body area ratios between cases and controls. DE, deep extensor; DF,
deep flexor; SE, superficial extensor; SF, superficial flexor.
Fig. 4 Error-bar chart demonstrating the comparison of various flexor–extensor cross-sectional
area ratios between cases and controls. DE, deep extensor; DF, deep flexor; SE, superficial
extensor; SF, superficial flexor; TE, total extensor; TF, total flexor.
Interobserver Variability
The agreement between the two observers for the measurement of the PSM areas ranged
from substantial to almost perfect (weighted kappa coefficients: 0.78 for DF, 0.82
for SF, 0.76 for DE, and 0.85 for SE).
Discussion
Biomechanical Abnormalities in HD
The repeated flexion-induced microtrauma in the lower cervical cord in HD patients
has been commonly attributed to factors like an inherently tight dura,[13 ] growth disparity between the spinal column and the cord,[14 ] flexion-induced venous stagnation,[15 ] and anterior osteophytes.[13 ] The role of spinal biomechanics in the pathogenesis of HD has been under-explored.
Some studies have described an exaggerated range of vertebral flexion in these patients.[16 ]
[17 ] This flexional hypermobility is postulated to enhance the anterior displacement
of the dural sac and aggravate cord injury and atrophy.[18 ] HD patients have also been noted to have loss of cervical lordosis and impaired
sagittal balance parameters that improve after anterior cervical fixation procedures.[19 ] Instability itself has also been postulated to contribute to the disease process
on the basis of accompanying findings like abnormal cervical curvatures and the presence
of osteophytes[20 ] in these patients. The demonstration of altered cervical PSM morphology in our study
reflects yet another biomechanical perspective to the disorder.
Paraspinal Muscle Changes in HD
MRI-based measurement of CSAs of muscles is a robust proxy measure of muscle strength.[21 ] It has been found to be a reliable indicator of muscle atrophy and weakness, with
excellent intra- and interobserver agreement.[22 ]
[23 ]
[24 ] With respect to HD, Li et al[7 ] had analyzed CSAs of cervical PSMs at two levels and had concluded that HD patients
have smaller muscle areas compared with controls, and that there is an imbalance between
the SF and SE CSAs. These findings were taken to reflect an underlying biomechanical
instability. The findings of the study were, however, not generalizable to the entire
subaxial spine as the measurements were restricted to limited levels in the lower
cervical spine where PSM changes could have occurred as a direct consequence of anterior
horn cell dysfunction.
The comprehensive PSM area measurements in our study demonstrate that the paraspinal
morphometric alterations in HD extend to all levels of the subaxial cervical spine.
Li et al's[7 ] conclusion of all cervical PSMs being uniformly smaller in HD patients is not borne
out by our data. We found that while the subaxial extensor muscles in HD patients
are indeed smaller, their flexor PSMs are uniformly larger than those of controls.
This was found to be true for both the superficial and deep muscle groups. Interestingly,
the flexor–extensor muscle disparity was noted to occur at all subaxial levels in
a condition that primarily affects the C7-T1 myotomes, indicating that the biomechanical
derangements in HD are pan-cervical.
Muscle Group Co-contraction and Disparity
While the superficial cervical muscles are understood to assist predominantly in voluntary
neck movements, the deep muscles control segmental motion and maintain cervical alignment.
However, all motions of the cervical spine occur as a result of a delicate interplay
of the flexors and extensors of the superficial and deep muscle groups. This co-contraction
of different muscle groups stiffens the spine and contributes to its static and dynamic
stability.[25 ]
[26 ]
[27 ]
[28 ] Impaired co-contraction of the flexors and extensor muscles under different movement
conditions has been linked to the occurrence of various neck disorders[29 ]
[30 ] including HD.[7 ]
The possibility of a flexor–extensor disparity in HD was previously considered on
the basis of a large SF/SE CSA ratio that was taken to imply a more pronounced hypotrophy
in the extensors than the flexors.[7 ] Though other studies in HD have not directly alluded to this disparity, some did
report loss of cervical lordosis[31 ]
[32 ]—a finding that intrinsically suggests the existence of weak extensors.[33 ] Our study confirms the existence of a flexor–extensor PSM disparity across all subaxial
levels in HD. The larger flexor–extensor ratios in our study are not due to varying
degrees of a generalized hypotrophy as reported previously,[7 ] but due to larger flexors and smaller extensors than those of the normal population.
Implications of This Study
This study has potential implications in modifying the conservative treatment protocol
for HD, currently restricted to long-term immobilization with a cervical collar. Patient
compliance with this modality is poor in view of the prescribed requirement of continuous
collar-wear for a minimum of 2 years.[34 ]
[35 ]
[36 ] This approach also carries the risk of causing disuse PSM atrophy in spines that
are biomechanically unsound to begin with.
Our data has established a generalized subaxial cervical flexor–extensor disparity
in HD patients. With the current analysis, it is difficult to comment on whether the
larger flexors and smaller extensors in HD patients have a cause or effect relationship
with the disease and its proposed pathogenetic mechanisms. Either way, it would be
worth exploring whether the pronounced flexor–extensor disparity in these patients
is modifiable by approaches such as extensor muscle strengthening, muscle co-contraction
retraining, and postural re-education.[37 ] Such targeted physiotherapy protocols could unfold a more physiological method of
managing HD and potentially help in arresting the clinical course of the disease.
Limitations
This study is limited by the inherent disadvantages of a retrospective study, and
prospective trials will help in further elucidating the interactions between PSMs
and the clinical course of HD. Though MRI-based muscle CSA measurement is an appropriate
marker of muscle strength, direct electromyographic evaluation of the PSMs could possibly
strengthen the inferences of our study. Intraobserver agreement data for CSA measurement
was not available in our study. To further strengthen the findings of the current
study, it may be useful to analyze changes in muscle areas and a possible reduction
of the flexor–extensor PSM disparity following conservative or surgical management
of HD.
Conclusion
Our study indicates that patients with HD have morphometric alterations at all levels
of their subaxial cervical paraspinal musculature. These patients have abnormally
large flexors and small extensors compared with controls. This flexor–extensor muscle
disparity could be utilized as a potentially modifiable factor in the management of
the disease.
