Keywords
ossification of ligamentum flavum - surgical treatment - prognostic factors
Ossification of ligamentum flavum (OLF), though rare, is being increasingly recognized
as a cause of thoracic myelopathy. Earlier, OLF has been reported almost exclusively
in East Asian countries, particularly in Japan and Korea. Studies of OLF from other
regions, such as India,[1]
[2]
[3]
[4]
[5] Middle East,[6] and West including Caribbean,[7]
[8]
[9] are increasingly being reported. This is one of the largest study reported, so far,
in literature. Adults in the age group of 30 to 60 years are affected most frequently,
and the disease has a strong predilection for the lower thoracic spine (T9–T12). Because
of the low prevalence of this disorder, small patient population, and the rarity of
large clinical reports, the treatment guidelines and surgical prognosis in patients
with this disease remain unclear. In the present study, we retrospectively analyzed
the clinical features of patients with symptomatic thoracic OLF treated surgically
at our institute.
Clinical Materials and Methods
Data were collected from medical records to define the preoperative clinical features
of OLF myelopathy patients including age and sex, signs and symptoms, and duration
of symptoms. Neurological status was assessed preoperatively and postoperatively and
surgical outcome was determined after the OLF surgery. We used Nurick system[10] for evaluation of patients. We evaluated the correlation between outcome and age,
gender, level and number of segments affected by OLF, preoperative duration of symptoms,
preoperative neurological status, radiological findings, and the presence of dural
adhesions observed intraoperatively.
Patient Characteristics
One hundred and forty-three patients with thoracic OLF underwent surgery between 1989
and 2010 for the treatment of thoracic OLF in our institution. There were 87 male
and 56 female patients who ranged in age from 15 to 78 years (mean, 45 years). The
symptoms had been present for few weeks to 4 years prior to treatment. A total of
101 patients (70.6%) had duration of symptoms ≥1 year. A majority of our patients
(79.02%) were able to walk only with assistance or were wheel chair-bound/bed-ridden
(Nurick grade 4 and 5) indicating advanced disease at presentation. The diagnosis
was established after thorough neurological examination and subsequent imaging studies
including myelogram, computed tomography (CT), and magnetic resonance imaging (MRI).
All patients were evaluated and underwent surgical intervention. The postoperative
follow-up duration ranged from 12 months to 18 years (mean, 60 months).
Demographics of patients are shown in [Figs. 1] and [2]. Nearly 80% of patients were in the age group of 30 and 60 years which is the most
commonly affected age group. Most common presenting symptoms included difficulty in
walking (97%), numbness (93%), backache (85%), and urinary disturbances in 78 (54.54%)
patients. Spasticity and lower limb weakness was present in nearly all patients (>97%).
Two patients had hypotonia. ([Table 1]) The lesions characteristics of patient are shown in [Fig. 3], [Table 2], and [Figs. 4]
[5]
[6].
Fig. 1 Age distribution: X axis represents age groups and number of patients is represented
along Y axis.
Fig. 2 Sex ratio of patients.
Table 1
Clinical characteristics
|
No. of patients
|
Percentage (%)
|
|
Symptoms
|
|
Walking difficulty
|
139
|
97.20
|
|
Numbness of lower limbs
|
134
|
93.70
|
|
Backache
|
122
|
85.31
|
|
Sphincter disturbance
|
78
|
54.54
|
|
Signs
|
|
Spasticity
|
141
|
98.6
|
|
Hypotonia
|
2
|
1.39
|
|
Paraparesis/paraplegia
|
139
|
97.20
|
|
Sensory disturbance
|
135
|
94.40
|
|
Sphincter disturbances
|
78
|
54.54
|
Fig. 3 Shows lesion distribution in thoracic spine.
Table 2
Lesion characteristics
|
Level of involvement
|
No. of patients
|
Percentage (%)
|
|
Long segment involvement
(laminectomy ≥ 7 levels)
|
4
|
2.8
|
|
Two separate levels
|
8
|
5.6
|
|
Concomitant cervical + dorsal involvement[a]
|
21
|
14.68
|
|
Concomitant dorsal + lumbar involvement[a]
|
44
|
30.76
|
a Note the significant number of patients with coexisting cervical and lumbar ligamentum
flavum along with the thoracic involvement. This makes evaluation of whole spine necessary
in a case with symptomatic ligamentum flavum of one region of spine.
Fig. 4 (a-c) Magnetic resonance appearance of ligamentum flavum showing hypointense excrescences
into spinal canal projecting from posterior aspect. (a) Upper thoracic ligamentum flavum, showing round type morphology, (b) middle thoracic ligamentum flavum, showing a beak type morphology, and (c) lower thoracic, which is the commonest site of involvement by ligamentum flavum.
Fig. 5 Sagittal T2-magnetic resonance images showing concomitant involvement of lumbar (a) and cervical spine (b) along with thoracic ligamentum flavum—this is a common finding emphasizing the fact
that whole spine examination is necessary in patients with ligamentum flavum of one
region of spine.
Fig. 6 (a–d) Different types of ligamentum flavum on axial computed tomography: appearing as
hyperdense nodules projecting into the spinal canal. (a) Lateral type, (b) enlarged type, (c) fused type, and (d) tuberous type.
Imaging
Because of unavailability of MRI at our center during initial period, myelogram was
performed in 24 patients and CT alone was done in 2 cases. MRI study of whole spine
was performed in 117 patients (81.8%) with or without CT scans. Cord hyperintensities
on T2 images were present in 33 patients preoperatively.
Surgical Procedure
Surgical treatment is indicated for symptomatic OLF. The goal of surgical treatment
was a complete cord as well as nerve root decompression at the involved level, and
a decompressive laminectomy and resection of the OLF was performed in all cases. We
used drill for performing the laminectomy and thinning of OLF. The range of laminectomy
included one lamina superior and inferior to the diseased segments in addition to
the medial one-third facetectomy through the diseased segment. If ossified dura mater
was present, it was resected trying to preserve the arachnoid mater ([Fig. 7]). If cerebrospinal fluid (CSF) leakage was found, the dura mater was repaired. Five
patients underwent second surgery for symptomatic OLF at other levels later. No fusion
was performed in any of our patients. We not only decompress the cord, but also look
for nerve roots in intervertebral foramina and if needed, they were decompressed also.
Fig. 7 Intraoperative photograph showing the dural calcification in a case of ligamentum
flavum.
Results
Post-op Evaluation
We evaluated our patients using Nurick system[10] ([Table 3]). Eighty-six patients (60.13%) were able to do their job with mild neurological
deficit such as spasticity and impairment of joint position sense. Twenty-eight patients
(19.58%) were back to some job with limitation but were independent for their routine
work. Sixteen patients (11.18%) improved but still need help for their routine work
while thirteen patients (9.09%) were bed-ridden at last follow-up. ([Table 4]).
Table 3
Nurick grades
|
Grade 0
|
Root signs and symptoms, no sign of spinal cord involvement
|
|
Grade 1
|
Signs of spinal cord involvement, but no difficulty walking
|
|
Grade 2
|
Slight difficulty walking, which does not prevent full-time employment
|
|
Grade 3
|
Gait abnormality prevents full employment, but doesn't require someone else's help
for walking
|
|
Grade 4
|
Able to walk only with assistance/walker
|
|
Grade 5
|
Chair-bound or bed-ridden
|
Table 4
Outcome
|
Outcome
|
No. of patients
|
Percentage (%)
|
|
Good:
Improvement by ≥ 3 grades
Nurick grade at last follow-up 0–2
|
86
|
60.13
|
|
Fair:
Improvement by 2 grades/Nurick grade at last follow-up 3
|
28
|
19.58
|
|
Improved:
Improvement by 1 grade/Nurick grade at last follow-up 4
|
16
|
11.18
|
|
Poor:
no Improvement/neurological deterioration
Nurick grade at last follow-up 5
|
13
|
9.09
|
Complications
Postoperatively, seven cases (4.89%) had infection and eleven cases (7.69%) had CSF
leak. Twelve (8.39%) cases deteriorated neurologically. Five (3.4%) cases required
second surgery for ossification at other level during follow-up. The mean follow-up
was 5 years. Statistical analysis was done using chi-square test and an odds ratio
(OR) was calculated.
Prognostic Factors
We evaluated the relationship of various preoperative parameters with surgical outcomes
(good outcome as postoperative Nurick grade ≤ 3 and poor as Nurick grade 4–5) and
found that out of these only the short duration of symptoms and preoperative neurological
status reached statistical significant value ([Table 5]). There was no relation of sex, age, level of OLF, number of OLF affected segments,
and type of OLF with surgical outcome. But we believe that the following factors were
also important in predicting the surgical outcome, though not reaching a statistically
significant value: degree of spinal cord compression, absence of proximal stenosis,
intramedullary hyperintensity on T2 images, intraoperative cord manipulations, and
degree of decompression achieved.
Table 5
Correlation between preoperative features with surgical outcome
|
Parameter
|
No. of patients
|
Functional outcome
|
x2
test
|
p
|
OR
(95% CI)
|
p
|
|
Good
|
Poor
|
|
Age
|
≤60
|
128
|
103
|
25
|
0.423
|
0.52
|
1.498
(0.440–5.100)
|
0.52
|
|
>60
|
15
|
11
|
4
|
|
Sex
|
Male
|
87
|
67
|
20
|
1.008
|
0.31
|
0.641
(0.269–1.532)
|
0.32
|
|
Female
|
56
|
47
|
9
|
|
Levels involved
|
<4
|
110
|
87
|
23
|
0.117
|
0.73
|
0.841
(0.310–2.278)
|
0.73
|
|
≥4
|
33
|
27
|
6
|
|
Pre-op duration of symptoms
|
<1 year
|
42
|
38
|
4
|
4.255
|
0.039[a]
|
3.125
(1.015–9.626)
|
0.047[a]
|
|
≥1 year
|
101
|
76
|
25
|
|
Cordhyper intensity
(n = 117)
|
Present
|
33
|
25
|
8
|
2.164
|
0.14
|
0.471
(0.170–1.303)
|
0.12
|
|
Absent
|
84
|
73
|
11
|
|
Dural adhesions
|
Present
|
83
|
65
|
18
|
0.242
|
0.62
|
0.811
(0.351–1.872)
|
0.62
|
|
Absent
|
60
|
49
|
11
|
|
Pre-op Nurick grade
|
≤3
|
30
|
28
|
2
|
4.352
|
0.037[a]
|
4.395
(0.982–19.665)
|
0.043[a]
|
|
4–5
|
113
|
86
|
27
|
a Shows statistically significant values.
Discussion
OLF is no longer a condition seen only in East Asia, but is present worldwide. Still
it is a relatively rare entity with unknown origin, insufficient epidemiological data,
and no guidelines or standards for its treatment. The ligamentum flavum is a yellow
elastic ligament extending from the second cervical vertebra to first piece of sacrum.
The ligamentum flavum bilaterally has two components, medially interlaminar portion
and laterally capsular portion.[5] It is believed that the hypertrophy of ligamentum flavum precedes the development
of OLF. The early ossification is usually found on capsular side of ligament in front
of the facet joints. Then it expands on the outer surface of ligament anteriorly toward
the spinal cord gradually giving rise to two paramedian nodules within the spinal
canal that compresses the posterior portion of the spinal cord (
[Fig. 6a–d]). Rarely, the lateral capsular outgrowth impinges on an existing nerve root resulting
in foraminal stenosis and radiculopathy. Both masses are usually connected by a film
of elastic fibers which subsequently also ossifies on its epidural surface and occupies
the spinal canal in the form of a central mass. This further contributes to cord compression
as it thickens and extends in a caudocranial direction along the posterior border
of spinal canal in a linear fashion as a beak-like outgrowth.[3] OLF is the result of a metaplastic process in which enchondral ossification leads
to lamellar bone formation[12]
[13] ([Fig. 8]) It is reported to occur most commonly in thoracic, then lumbar, and rarely in cervical
spine.[14] ([Fig. 5a, b])
Fig. 8 Histopathological photomicrograph of an ligamentum flavum specimen showing fibro
collagenous tissue (pink) produced due to elastic fiber breakdown with calcium deposits
(blue; hematoxylin and eosin stain).
Thoracic OLF most commonly involves the lower thoracic spine vertebrae between T9
to T12,[13]
[15]
[16]
[17] where greater mobility and vulnerability (due to spinal motion) may result in frequent
mechanical injury along with a unique orientation of zygapophyseal joints that contribute
to increased rotatory instability and micro mobility.[18] In the present study, a similar distribution was observed. OLF can occur in upper
and mid thoracic level, ([Fig. 4a-c]) and causes can be multifactorial. Mechanical stress involving constraints on the
articular processes during rotation might be a critical factor.[17] The elastic fiber breakdown and increase in the number of collagen fiber in ligamentum
flavum has been reported to exhibit increased affinity for calcium and promote the
formation of calcium deposits, chondrometaplasia, and enchondral ossification. The
factors released during repair of damaged ligament from mechanical stress for example,
fibronectin,[19] bone morphogenic protein[20] and transforming growth factor-β[21]
[22] may play a role in genesis of OLF. OLF is frequently associated with narrowing of
neural canal, ossification of the posterior longitudinal ligament, or ossification
of other ligaments and symptomatic thoracic disc herniations. A high incidence of
coexisting spinal disorders in patients with thoracic OLF makes the examination of
whole spine necessary.[23]
It has been documented that incidence of thoracic OLF is higher in the patients with
diffuse idiopathic skeletal hyperostosis,[24] fluorosis,[2]
[25]
[26] diabetes,[27] and ankylosing spondylosis. In our study, fluorosis was found in 32 (22.37%) patients,
which can be correlated as an etiological factor in pathogenesis of OLF. It is reported
that flouride (F) can activate intracellular c-AMP, causing a significant increase
of intracellular calcium (Ca)[28] and could stimulate the activation and proliferation of osteoblast-like cells with
enhanced expression of mRNA and the proteins c-fos and c-jun.[26] The retention of Ca in chondrocytes caused by F leads to degeneration and necrosis
of chondrocytes followed by chondral matrix calcification and enchondral ossification.
These findings suggest that Ca and F might play an important role in pathogenesis
of OLF.[29] In our study, one characteristic finding noted in fluorosis-induced OLF was involvement
of large number of spinal segments.
The symptoms of OLF are secondary to stenosis of the spinal canal. The most common
clinical presentation consists of progressive thoracic myelopathy, resulting in spastic
paraparesis, with or without sphincter dysfunction. Walking impairment is generally
reported as the main complaint. OLF can also lead to loss of balance due to posterior
column dysfunction.[5]
[15] Sensory, bladder, and bowel involvement are not rare in the later stage.[30]
Before the early 1990s when MR imaging was not commonly used for neurological diagnosis,
myelography was considered the best diagnostic tool to detect OLF-induced thoracic
cord compression. In our study also, myelogram was performed in initial period. Myelography
demonstrates the level and posterior location of the compressive lesion but not its
nature. Axial CT scan reveals the contour and density of ossificatory change in ligamentum
flavum. A CT classification of OLF has been devised by Sato et al[31] dividing the OLF into five (lateral, extended, enlarged, fused, and tuberous) types
([Fig. 6a–d]). The T2-weighted sagittal imaging by MRI is modality of choice for the screening
of the longitudinal extent of OLF. On MR imaging, OLF can be recognized by the presence
of a triangular or hemispheric area of low signal intensity on T1- and T2-weighted
images in the posterior margin of the canal.[32] ([Table 4]; [Figs. 4], [5]) OLF has been “missed” on MR and later diagnosed by CT at some instances,[5]
[33] (as the null signal of ossification can be difficult to detect on MR imaging, unlike
hyperdense ossified lesion detected on CT imaging). It is therefore useful to combine
CT and MR images.
Conservative treatment is not effective in preventing disease progression and neurological
deterioration. Choice of a surgical procedure may be guided by type of ligamental
ossification. Although several techniques have been used such as laminoplasty,[13] foraminotomy,[34] image-guided laminotomy,[35] enblock resection,[36] and laminectomy with fusion.[17] But due to advanced nature of disease at presentation (as indicated by clinical
symptoms/involvement of dura in large [58%] number of our patients) we recommend that
laminectomy and complete resection of the OLF is the treatment of choice. There are
other reports in literature about using such method.[6]
[37]
[38] We routinely perform decompressive laminectomy along with medial one-third facetectomy
and resection of the OLF. The results show that this type of surgical method did not
affect the stability of the spine in patients with OLF and no additional internal
fixation and bone grafting was required. Despite good posterior decompression of thoracic
myelopathy due to OLF, recovery varies widely from 25 to 100%.[39] We had improvement in over 90% of our patients at their respective last follow-up.
Several peri- and postoperative complications have been reported in patients with
OLF-related myelopathy.[23] CSF leakage following disruption of dura is one of the major intraoperative complications.
In this series, 68 (47.5%) patients had a tear of dura mater, which was repaired.
Therefore, intraoperative manipulation should be gentle to avoid duramater injury
as much as possible, and once a dural tear is noticed, immediate repair should be
performed. Calcification of duramater was present in 83 (58.04%) of our patients,
indicating the advanced disease at presentation. One patient also had intradural calcification.
Postoperative CSF leak was observed in 11 of our patients (8.39%) and 7 (4.89%) had
localized infection which were managed conservatively with satisfactory outcome.
Neurological deterioration can occur immediately after surgery because of unintended
intraoperative spinal cord manipulation. Severe OLF results in spinal stenosis and
maximum cord compression. So, no instrument should be negotiated inside the spinal
canal that can cause irreversible spinal cord injury. High speed electric or pneumatic
drills, followed by thin footplate rongeurs (1 mm) should be used to avoid this complication.
Epidural hematoma is another complication that can cause postoperative spinal cord
compression and induce symptom aggravation.[15] No postoperative epidural hematoma occurred in our series. Some authors raise concern
about increased spinal kyphotic deformity after laminectomy that can cause late-onset
neurological deterioration or localized back pain.[15] We have not encountered any patient developing kyphotic deformity with long laminectomy
even after years of follow-up.
We found short duration of symptoms and preoperative neurological status as the most
important factors predictive of outcome of statistical significance. ([Table 5]) Under chronic compression the spinal cord exhibits destructive changes believed
to cause motor paresis: these include infarction, ischemic necrosis, and neuronal
loss or chromatolysis in gray matter as well as demyelination in white matter and
in posterior and lateral white columns.[30] It has been observed that apoptosis precedes the destructive changes in a spinal
cord which is under chronic compression, resulting in profound and irreversible neuronal
degeneration in posterolateral long tracts.[40] This might explain the persistence of residual spasticity at the follow-up, due
to irreversible changes within cord by thecal compression. Delay between onset of
initial symptoms and surgical decompression is believed to worsen these changes.[30] In other words, the short duration of symptoms, preoperatively, is a positive predictor
of outcome.[41]
A recent study evaluating the effect of intramedullary signal changes on recovery
suggested that the presence or absence of signal changes did not correlate with postoperative
recovery; but whenever present, intramedullary signal size (ISS) greater than 15 mm
significantly compromised recovery.[39] The results of our study match with other studies[39]
[42]; however, we did not evaluate the effect of ISS with recovery.
Conclusion
The pathogenesis of thoracic OLF is mainly due to localized mechanical stress on the
ligament and patients with fluorosis can have extensive disease. MRI is the investigation
of choice as it helps in screening the longitudinal extent of disease and the cord
compression. CT helps in identifying the type of OLF and type/extent of surgical resection.
Wide laminectomy with excision of OLF and calcified dura is the treatment of choice.
OLF is a rare cause for thoracic myelopathy and the surgical excision is the only
treatment modality for it.
