DEGENERATIVE DISC DISEASE

The social impact and suffering that results from back
pain cannot be overestimated. With a lifetime incidence of 60% to 80%,
low back pain (LBP) generates at least 15 million office visits per
year (23). In fact, low back complaints are the
leading compensable cause of injury in the workplace. In the United
States, $24 billion a year is spent on the evaluation and direct
management of patients with back pain. Indirect costs include work and
productivity losses, and they account for an additional $27 billion
annually (28).

Typically, back pain has a benign course. Within 3
months of the onset of symptoms, 95% of patients return to their
previous employment. The 5% of patients with residual symptoms after 3
months, however, incur 85% of the costs. Moreover, the probability of
returning to work falls with increasing duration of disability. After 2
years off work, less than 2% will return. Therefore, the early
detection of those patients in whom LBP is more likely to become
chronic would be of tremendous clinical and social benefit.

Adult LBP can arbitrarily be divided into five categories:

Before focusing on the spine and its related structures,
consider other potential sources of referred pain. Included in the
first LBP group are intra-abdominal and retroperitoneal pathologies
such as abdominal aortic aneurysm and endometriosis. In groups 2
through 4, symptoms more readily correlate with evident spinal
pathology; as a result, treatment of these patients is satisfying and,
overall, associated with good results.

The fifth group, patients with mechanical LBP only,
includes several benign conditions, most likely representing
ligamentous and muscular strain, mechanical stress from poor posture,
or facet joint irritation. More chronic and disabling degenerative disc
conditions, however, are included as well. In these patients, uncertain
identification of the pain generator is associated with vague
diagnostic groupings and failed surgical management. Included among
these is DDD.

The multiple and synonymous terminology reflects an
incomplete understanding of the pathophysiology and natural history of
DDD. Moreover, because the “pathoanatomic” changes noted in DDD do not
qualitatively differ from those of normal aging, the appropriateness of
the appellation disease is intensely debated.

It is theorized that in DDD, pain arises in the disc.
Historically, while the lumbar facets, posterior longitudinal ligament,
dura, dorsal root ganglion, and myofascial structures of the lumbar
spine have been recognized as pain-generating structures, the disc was
felt to be aneural. More recently, however, nociceptive fibers have
been identified in the outer annulus (22,30,47).

In one study, pain was generated by applying pressure to the disc during operations under local anesthesia (15). In another study, back pain resolved with procaine injection into the disc (75). Kuslich et al. (47)
used a progressive local anesthesic technique to gauge pain response in
different tissues in 193 consecutive laminectomies. Stimulation via
blunt probe or unipolar electrocautery on the facet cartilage and
synovium never caused pain. Also, no pain followed stimulation of the
lamina, spinous process, ligamentum flavum, lumbar fascia, and
uncompressed roots. Stimulation of the facet capsule, however, was
associated with sharp, localized pain in 30%. This pain did not match
the patient’s preoperative symptoms. Stimulation of a compressed root
resulted in sciatic pain in 79%. Finally, LBP similar to preoperative
symptoms was noted in 70% of patients after stimulation of the
posterior annulus or posterior longitudinal ligament (PLL). Local
anesthetic injection obliterated the pain.

Nociceptive free nerve endings of the recurrent
sinuvertebral nerve most likely carry these impulses. The sinuvertebral
nerve, first described by Von Luschka, consists of a postganglionic
derivative of rami communicantes that branches into segments. These
segments ascend and descend into one or more adjacent levels (30,78) (Fig. 145.1).
The branches accompany the venous plexus into the vertebral endplates
and terminate in nociceptive free nerve endings in both the PLL and
outer lamina of the annulus (51). In
histometric studies, the outer third of the annulus has been found to
contain nerve endings with nociceptive neurotransmitters [substance P,
calcitonin, vasoactive intestinal peptide (VIP)] (54). Further, nociceptive fibers may grow into diseased discs (22). Theoretically, a small tear of the outer annulus may cause pain, even with a normal nucleus pulposus.

Disc degeneration may indirectly stimulate pain
receptors elsewhere. Disc height collapse may produce nociceptive
signals from the facet mechanoreceptors by abnormal loading. Such
mechanical derangement has been associated with abnormal intervertebral
motion and has been termed lumbar segmental instability.

A degenerated disc may also indirectly cause pain by
extrusion of nuclear material, a source of neural irritation and
inflammation (54). The potential sources of pain in the lumbar spine are illustrated in Figure 145.2.

During an individual’s lifetime, intervertebral disc
composition changes greatly. Given that certain degenerative phenomena
may lead to pain, it is necessary to have a clear understanding of what
changes constitute normal degeneration.

At birth, the disc surface area is 50% nucleus pulposus
(NP) and 50% annulus. The notochordal cells of the NP are gradually
replaced by chondrocytes throughout the early teenage years. This
replacement is associated with annular thickening (Fig. 145.3). The demarcation between

the annulus and the nucleus becomes less distinct. The older NP has a
higher collagen content with more structured fibers. In these fibers,
the ratio of type II to type I collagen increases (51).

Proteoglycan metabolism changes with age as well.
Chondroitin-4-sulfate and chondroitin-6-sulfate concentrations
decrease, and the ratio of keratin sulfate to chondroitin sulfate
increases. Keratin sulfate has a smaller hydrophilic potential and a
reduced tendency to form stable aggregates with hyaluronic acid.
Dehydration, in turn, decreases the disc’s resistance to axial loading (20).

As desiccation continues, clefts can be identified that
originate in the central portion of the dehydrated NP. One hypothesis
holds that these clefts eventually migrate toward the peripheral
annulus and endplate and cause tears. Annular tears are classified by
Vernon-Roberts (20) as peripheral,
circumferential, and radiating. Circumferential and radiating tears are
associated with degenerative changes in the endplate and NP.

With maturity, the vascular supply to the disc is
obliterated. In adults, the disc is the largest avascular structure in
the body. Thereafter, disc nutrition requires diffusion through the
vertebral endplates (80%) and outer annulus (20%) (25).

Some authors have differentiated age related changes
from pathological degeneration. One study found that altered collagen
staining patterns and increases in lipofuscin and amyloid can be used
to differentiate aged from degenerated discs (20).
Further, this cycle of degeneration may be self-promoting, that is,
small annular tears lead to further nuclear degeneration in animal
models (47,54). Vascular ingrowth may also mark degenerated and herniated discal tissues (41).

At present, there is no clear boundary between these
aging changes and disc degeneration. While some authors claim a
quantitative if not qualitative difference between aged and degenerated
discs, all discs degenerate with age. Miller et al. (55) reported evidence of disc degeneration by the age of 50 in 90% of 600 autopsy specimens. Holt (39) found evidence of disc degeneration on the plain radiographs of 80% of adults studied, although 53% had no history of LBP.

While it appears that all intervertebral discs
degenerate with age, the degree and rate of this degeneration vary
significantly from individual to individual. The underlying reasons for
this variability are only partly known.

The Hirsh theory of disc degeneration holds that
insufficient nutrition, impaired waste transport, and traumatic
mechanical factors combine with a genetic and hormonal proclivity to
cause desiccation and annular tearing. Severe degeneration is
associated with increased lactate metabolism, decreased pH,
accumulation of proteolytic enzymes, and chondrocyte necrosis (Fig. 145.4) (37).

Various clinical factors have been implicated in
precocious degeneration. Smoking and a familial tendency toward
degeneration have long been established as contributing factors (68). Twins demonstrate similar degeneration patterns (3). In one study of 15-year-olds, LBP, decreased activity, and decreased spinal range of motion predicted later DDD (66). Anatomically, no association between DDD and facet tropism can be demonstrated (6). The endplate irregularities of thoracolumbar Scheuermann’s disease, however, may be related to DDD (35).

If disc degeneration is a programmed, physiologic
phenomenon influenced by heredity and environment, it is incumbent upon
the clinician to ascertain when these changes constitute a disease.
Benign age-related phenomena have been differentiated from pathologic
phenomena on the basis of three factors: impaired function, structural
changes, and an association with pain.

Yong-Hing and Kirkaldy-Willis (86) described a three-joint complex in the spinal column, with the disc anteriorly

and the two facet joints posteriorly. They theorized that benign
microscopic alterations progress to pathologic degeneration in stages.
In this way, circumferential annular tears progress to radial tears.
Radial tears, in turn, engender further disc degeneration or frank disc
herniation. The ensuing loss of disc height alters facet joint
mechanics, and facet cartilage disruption or destruction may take
place. Coincidentally, the decrease in the intervertebral height causes
buckling of the ligamentum flavum and facet overriding or enlargement.
These changes may, singly or in combination, cause narrowing of the
neuroforaminal and central canals. So, while disc degeneration
manifests as mechanical LBP in some patients, others will experience
neurologic claudication or radiculopathy from frank disc herniation (Fig. 145.5, Fig. 145.6).

Yong-Hing and Kirkaldy-Willis (86) then proposed three stages of degeneration of the intervertebral disc.

Vernon-Roberts (20) showed that
many of these changes are present in nearly all middle-aged people and
are not necessarily associated with back pain. Thus, the issue of when
degenerative changes represent a disease remains unresolved. While data
are lacking, it has been suggested that younger patients with
relatively precocious disc degeneration do not tolerate these changes
as well as older adults. Whether this perceived difference stems from
higher functional demands or from a subtle difference in disc mechanics
is not known. It is reasonable, however, to identify DDD as a chronic,
mechanical LBP syndrome associated with changes in the structural and functional integrity of one or more intervertebral discs.

At present, DDD can be characterized by one of several associated terms, listed in Table 145.1.
While some authors use these terms to refer to the same global
discogenic pain syndrome, others view them as a means to differentiate
among subgroups of patients. The divisiveness and misapplication of
nomenclature further confuses any evidence-based appraisal of DDD, and
its incidence, pathophysiology, and natural history.

Three subtypes may be identified as points in a continuum. First, internal disc disruption
(IDD) refers to a painful annular tear in the absence of bony changes
or disc height loss. As such, IDD must not be confused with disc
protrusions, which are normal hydrodynamic findings (5).

Next, lumbar spondylosis (LS) refers to mechanical back pain in association with disc height loss, sclerosis,

and osteophyte formation. LS is the most frequently described form of
DDD, and it is the entity most authors are describing when they use the
more global term DDD.

Finally, lumbar segment instability
(LSI), which represents a progressive relaxation of facet capsules and
ligamentous restraints, occurs in the context of chronically
compromised disc biomechanics. White and Panjabi (83)
define lumbar instability as more than 4.5 mm of translation, 15° to
25° of angular motion between adjacent segments on flexion–extension
radiographs, or both (Fig. 145.7). It is not
clear, however, that the chronic increase in intervertebral motion seen
in degenerative lumbar diseases may be mechanically equated with
traumatic spinal instability. Therefore, some authors classify this
abnormality with degenerative spondylolisthesis and degenerative
scoliosis, which have a similar pathogenesis. Others feel that all the
subtypes of DDD represent a painful “microinstability” of the motion
segment (48).

The critical question of why most degenerated discs are
asymptomatic remains to be answered. Even in those patients with
degenerative changes, severe back pain, and positive discography, pain
may spontaneously improve. In one series, 25 patients who had not had
surgery were positive on a single-level discogram. When they were
evaluated after an average of 4.9 years, 68% had improvement without
surgery (70). Of the 32% who were unimproved,
66.7% had an underlying psychiatric diagnosis. In the absence of a more
complete understanding of the natural history of this disorder,
appropriate evaluation of surgical outcomes is extremely difficult.

In summary, while knowledge of the benign degeneration
of the aging spine continues to grow, surprisingly little is known
about disc degeneration as a disease process (Table 145.2) (78).

When evaluating a patient with chronic LBP, DDD remains
a diagnosis of exclusion. A thorough history and physical examination
are mandatory.

In patients with “red flags” such as very young or old
age, nonmechanical pain, constitutional symptoms, and trauma, perform a
thorough radiologic and serologic evaluation for infection, tumor, and
fractures (Table 145.3). Other important
considerations include intra-abdominal and intrapelvic pathology.
Posterior penetrating ulcers, pancreatitis, renal disease, abdominal
aortic aneurysm, and endometriosis are all known causes of severe,
referred pain to the back.

Pay careful attention to historical clues as well as to
significant social and psychological issues. Ask specific questions
regarding drug and alcohol intake, mood, sleep disturbance, pending
litigation, and job satisfaction.

In patients with pain of degenerative etiology, first
exclude those with radicular or myelopathic signs and symptoms.
Similarly, the evaluation of patients with significant thoracolumbar
deformities such as scoliosis, hyperlordosis, and kyphosis is
considered elsewhere (see Chapter 153, Chapter 155, Chapter 156, Chapter 159, Chapter 160 and Chapter 161).
For example, spondylolisthesis, another disorder of genetic and
environmental stress, is the most common structural abnormality in the
adult spine (see Chapter 162). Spondylolisthesis is related to LBP in 5% of the population (21).

The evaluation of the remaining group of patients with
isolated, mechanical LBP is challenging. The difficulty in identifying
a specific pain generator accounts for the fact that only 15% of
patients are given a definitive diagnosis. Physically and
radiographically abnormal structures may not cause symptoms. On the
other hand, grossly and radiographically normal structures may be
associated with severe pain in certain individuals.

Muscular etiologies account for the vast majority of
patients reporting acute LBP. Patients with myofascial pain are
especially difficult to distinguish from those with “discogenic pain”
when radiographic signs of disc degeneration are present. However,
myofascial pain tends to have an acute onset and a relatively brief
duration. The pain is localized to a specific paraspinal area, and
muscle spasm is evident. In most cases, while some DDD patients
identify a specific, traumatic event (such as bending, lifting, or
twisting) with symptom onset, their pain is midline and does not
resolve but rather worsens over time.

The pain described by patients with DDD is mechanical in
that it is aggravated by activity, particularly flexion. Relative rest
temporarily ameliorates symptoms. Patients with lumbar segmental
instability may complain of a catch with flexion and extension (Fig. 145.8).
DDD patients may also report having difficulty when getting up from a
chair. In an attempt to splint the back, some will use upper extremity
leverage, pressing their arms against their thighs, when arising.

Pain related to DDD may predate or postdate an episode
of sciatica. Often, DDD patients report having had prior discectomy or
chymopapain injections (20). Further, the disc collapse associated with LS may result in foraminal

stenosis and mild radiculopathies (48).
Other patients followed for axial spinal pain will later present with
acute radicular complaints and imaging studies consistent with disc
herniation. Diagnosis and treatment for these patients is discussed in Chapter 144.

Degenerative discs may be associated with referred,
sclerotomal pain to the buttocks and posterior thigh. However, it is
very difficult or impossible to localize the symptomatic disc level on
the basis of history or physical exam alone.

The examination of DDD patients is generally
nonspecific. These patients exhibit no point tenderness or paraspinal
spasms. They often report pain or difficulty with flexion and rotation
maneuvers of the spine, however. Normal neurologic findings including
sensory, motor, and reflex exams are expected. Particular attention
must be paid to Waddell’s signs (Table 145.4) (79), which suggest psychological overlay.

Moreover, unexpected findings on physical examination
such as anal sphincter laxity or major muscle weakness should be
construed as red flags and investigated accordingly.

The ubiquity of painless LS limits the specificity of
plain radiography in DDD. In the absence of red flags, radiographs are
indicated only after a trial of symptomatic treatment. In patients
failing to improve with these modalities, begin radiologic assessment
with plain anteroposterior (AP) and lateral views of the lumbosacral
spine. Oblique views and a lateral L5–S1 cone-down view are often
helpful. The principal purpose of these studies in the early management
of mechanical back pain is to exclude spondylolisthesis and the less
benign entities mentioned previously.

Patients with internal disc derangement will have no
plain radiographic changes. The cardinal findings of LS are endplate
sclerosis and loss of disc space. Radiographs may also show a loss of
lordosis, subluxations, vacuum phenomenon, and osteophytes (Fig. 145.9). Radiography, however, can be misleading: Frymoyer et al. (24) showed that signs suggestive of disc degeneration were present in 90% of adults studied, whereas 53% had no history of LBP (Table 145.5).

It should be noted that the presence of nitrogen gas
bubbles (the vacuum phenomenon) in degenerative discs probably excludes
the diagnosis of discitis, as infection by gas-forming organisms are
exceedingly rare.

In patients with normal findings on static radiography,
obtain flexion–extension radiographs to exclude subtle instability
patterns. Relatively subtly increased intervertebral motion can be
associated with pain. In these patients, discography may be useful to
establish a pain generator. Some authors assign no clinical importance
to lumbar segmental motion (24,32).

Radionucleotide bone scans have only a minor role in the
evaluation of disc degeneration by excluding other suspected pathologic
processes, such as tumor, infection, or spondylolysis. Computed
tomography (CT) may demonstrate degenerative changes in the lumbar
spine, but is not particularly useful in the evaluation of patients
with DDD.

In the evaluation of DDD, magnetic resonance imaging
(MRI) is the most commonly employed imaging modality. In that MRI can
directly measure water content of the disc, it is the only imaging
technique that can detect biochemical changes in the nucleus (56,61).
The normal, hydrated NP has an increased proton signal on T2 images.
With increasing desiccation, this signal blends with that of the
surrounding annulus. With further degeneration, a dark, isointense
signal may be seen on T2 (Fig. 145.10, Fig. 145.11).
An increased T2 signal may be noted in some areas, where it is thought
to represent free fluid in annular tears and fissures (87). Aprill and Bogduk (1)
described a high-intensity zone (HIZ) representing a tear of the outer
annulus. In their study, these HIZ lesions were associated with
painful, concordant discography. Subsequent reports as to the
significance of these lesions have been mixed (64).

Aside from the discs themselves, changes in endplate morphology adjacent to degenerating discs may have clinical significance (56). These changes were first described by Modic et al. (57)
and can be divided into three types. Type I reflects an acute
disruption and fissuring of endplates, which leads to ingrowth of
vascularized fibrous tissue into the adjacent vertebral body marrow.
This tissue exhibits a diminished signal on T1 images and increased
signal on T2 images.

In chronic degeneration, the hematopoietic (red)
peridiscal marrow undergoes fatty degeneration. Here, a type II pattern
is exhibited with an increased T1 and an isointense or slightly
hyperintense T2 signal. While type II changes tend to remain stable,
type I changes have been shown to develop into type II. Type III
changes probably reflect extrinsic bone sclerosis as seen on plain
radiographs. Dense bone in the vertebral endplates yields a hypointense
signal on both T1 and T2 images (Table 145.6).

Just as the pathologic changes of disc degeneration are
likely to be a normal part of aging, the MRI findings associated with
the changes described by Modic and Ross (56) frequently do not correlate with LBP. In 1990, Boden et al. (5)
found that MRI evidence of degenerative discs was present in 34% of
patients 20–29 years old, and in 93% of patients 60–80 years old.

Although discography is controversial, it represents the
only provocative method available to assess patients with a possible
discogenic pain generator. In theory, fluid injected into the disc
increases endplate pressures. These transferred pressures may cause
pain (34). Abnormal pressure transmission may
account for the small subset of patients with normal MRI findings and
positive discography. Properly performed, discograms may
be able to directly identify a cause-and-effect relationship between
radiographic signs of degenerated discs and clinical symptoms of lumbar
pain.

Morphologic information is available from the discogram or a postdiscography CT (Table 145.7, Fig. 145.12).
Abnormal radiographic findings with leakage of dye were seen in 37% of
an asymptomatic population, so the study is positive only if the
radiographic changes correlate with a concordant pain response. A
concordant response requires replication of the patient’s usual pain
with injection at the degenerated level, and no pain at adjacent,
control levels (Table 145.8).

In one large series, Holt (38), having found that a high percentage of previously asymptomatic volunteers had

positive discography, considered the test unreliable. This study was later criticized by Simmons et al. (69),
who noted that injections without fluoroscopic guidance often
pressurize the sensitive annulus rather than the nucleus. Further, the
contrast material used, diatrizoate meglumine, has been found to be
irritating and painful. Later, the Holt study was repeated using modern
techniques by Walsh et al. (80), who injected a
nonionic, water-soluble contrast agent under fluoroscopic guidance. The
study was considered positive only if the disc was radiographically
abnormal and a concordant pain response reported. A false-positive rate
of 0% with a specificity of 100% was noted in 10 asymptomatic
volunteers.

Calhoun et al. (8) found that
89% of 137 patients with positive provocative discography had
significant and sustained benefit from fusion at the indicated level.
Of 20 patients fused without a positive discogram, only 52% enjoyed
postoperative pain relief.

Given the overall controversy surrounding discography,
it can be expected that recommendations regarding its role in patient
assessment also vary. Some report that a negative MRI may miss
clinically significant DDD (7). Others report that positive discography in the context of a negative MRI is associated with inferior results after fusion (27).

Schneiderman et al. (67) found a
high correlation between MRI and discographic findings. They reported
on 101 disc levels in 36 patients with LBP of longer than 2 months’
duration. In each patient, both MRI and discography were performed and
blindly reviewed by a neuroradiologist. MRI was accurate in predicting
discographic disc morphology 99% of the time. Only one disc level with
a normal MRI signal had an abnormal discogram. Of 49 levels with
decreased signal on MRI, only two were normal on discogram. The authors
found that concordant pain with discography was helpful in the
assessment of abnormal discs identified by MRI, but they felt that
discography was not indicated in the presence of a normal MRI. Simmons
et al. (69), on the other hand, found only a
55% correlation between the tests. They wrote that discography is the
only dynamic test available for disc evaluation and thus the only study
that can determine which abnormal discs are truly symptomatic.

Horton and Daftari (40) found
that, in many cases, MRI could not reliably predict or replace
discography. They divided the MRI signal of lumbar discs into dark,
white, and speckled patterns, and they characterized the posterior
annulus as flat, bulged, or torn. Most dark or torn discs demonstrated
positive discography, whereas white or flat discs were very likely to
be negative on discography. The intermediate MRI patterns had uncertain
correlation with discographic findings, however.

At present, we feel that discography should be performed
only as a preoperative test in psychologically normal patients with
positive MRI findings, and after aggressive,

nonsurgical
measures have failed. Discography is probably not warranted in patients
with single-level changes. In patients with multilevel or equivocal MRI
findings, we use a discogram to detect the symptomatic level.

Various braces and supports have been recommended to provide pain relief in DDD (14,60,65).
The underlying hypothesis is that bracing will simulate effects of
fusion by restricting segmental motion. In general, such bracing is not
justified, in that the braces most commonly recommended are unreliable
in restricting lumbar spinal motion (17,65). Adequate immobilization of the lower lumbosacral spine requires a pantaloon brace.

Ultimately, the treatment of DDD is directed only at
pain, the perception of which is quite variable. Depression and anxiety
are quite common in DDD patients and are associated with heightened
pain perception (77). Historically, it has been
difficult to establish whether the pain preceded the psychological
disturbance. However, in a series of 200 patients with chronic back
pain, 77% met the American Psychiatric Association’s Diagnostic and
Statistical Manual of mental disorders (DSM III-R) lifetime criteria
for psychiatric illness, and 59% met criteria for current, active
psychiatric illness (63). In this study, the
authors concluded that psychiatric illness (particularly anxiety and
substance abuse disorders) often preceded the onset of back pain.

Indications of somatic fixation on the Minnesota
Multiphasic Personality Inventory (MMPI) include abnormal elevations of
the hypochondriasis and hysteria scales, above that of the depression
scale. On this inventory, somatic fixation tends to be predictive of
poor surgical outcomes (84). Patients with low
hypochondriasis and hysteria scores had 90% good to excellent results 1
year after surgery, while patients with higher scores had only a 10%
rate of good to excellent results (84). Southwick and White (71) reported that patients in the latter category were more likely to have positive discograms at nondegenerated levels.

While certain radiologic findings are consistent with
disc degeneration, there are no findings pathognomonic for DDD. As in
any degenerative condition of the spine, begin the evaluation with a
complete history and physical exam. Pursue atypical pain and other red
flags vigorously. Then, assuming limited findings on physical exam,
commence further management with a rigorous nonoperative management
regimen.

Only patients failing this management after a 2-month
trial require plain radiographic evaluation. While the changes on plain
radiographs associated with disc degeneration do not necessarily confer
a diagnosis, use plain radiography to exclude other potentially serious
causes of pain. If these radiographs are negative, obtain
flexion–extension lateral radiographs to rule out subtle instability.

In all cases, pay careful attention to psychosocial
factors. A long period of evaluation and nonoperative management
affords the surgeon an opportunity to get to know the patient well.
Obtain an MMPI if there are any doubts as to the psychological profile.
A motivated, professionally satisfied patient is the ideal candidate
for further evaluation and possible surgical treatment.

The MRI is the study of choice in the evaluating the
intervertebral disc. In the context of unresponsive mechanical pain in
a psychologically normal individual, single-level degenerative changes
may warrant consideration of operative treatment. Perform multilevel
discography if several levels are involved or MRI findings are
equivocal. Consider surgery only for those patients demonstrating one
(possibly two) levels of concordant pain with no pain at control
levels. Discography is at present not indicated in patients with a
normal MRI.

Posterolateral spinal fusion was previously the gold standard for the surgical treatment of discogenic back pain (36,62).
The advantage of the posterior approach is that fusion can be performed
in the absence of the posterior element; the risk of injury to the
neural elements is low; and because the graft is placed away from
midline, there is less risk of iatrogenic spinal stenosis.

Reported results of uninstrumented fusion have been contradictory. Stauffer and Coventry (73)
reported 89% good results, and they achieved an 80% fusion rate by
radiographic criteria; there was a high correlation between successful
fusion and the clinical result. Dawson et al. (11)
reported a 92% rate of solid fusion, with a 70% to 90% clinical success
rate. Those undergoing a fusion above the lumbosacral motion segment
were found to have a 45% pseudarthrosis rate, however.

Others have noted poor results with fusion. Finlayson et al. (18)
performed a posterolateral fusion for DDD in 20 patients with
concordant discography. While 11 felt the operation was worthwhile,
only six had a good outcome as measured by impairment, disability, and
work status. Parker et al. (62) reported a
prospective, consecutive series of patients with discogenic LBP
undergoing posterolateral fusion. They observed only 39% good to
excellent results, with 13% fair and 48% poor results. Poor results
were associated with workers’ compensation status, pseudarthrosis, and
being out of work longer than 3 months Greenough et al. (29)
concluded that posterolateral fusion was an acceptable treatment for
discogenic back pain only in very carefully selected patients.

Pedicle screw instrumentation has been added to
posterolateral fusion procedures in an attempt to decrease
pseudarthrosis rates (Fig. 145.13, Fig. 145.14).
Data from fusion procedures for degenerative spondylolisthesis suggest
increased fusion rates but no effect on clinical outcome (19).
Instrumentation is associated with higher costs and complication rates.
One recent series noted 10% of patients had instrumentation-related
problems. Yet, biomechanical studies suggest that pedicle screw
constructs are superior in stabilizing the nonosteoporotic spine (31).
These constructs may confer immediate stability to motion segments and
allow an expedited postoperative recovery. This added stiffness and
faster recovery interval may be more important in the younger
population with DDD than in the older patient with degenerative
spondylolisthesis.

Pseudarthrosis rates in noninstrumented fusions range
from 35% to 68%. Instrumented fusions have pseudarthrosis rates
reported from 0% to 33%. Increased rates of 75% to 95% significant
clinical improvement are also reported in patients undergoing
instrumented fusions, versus 59% to 70% in those fused without
instrumentation (3,88). In patients undergoing fusion for discogenic pain, solid fusion is associated with increased return-to-work rates as well (11).
Posterolateral fusions undertaken to treat DDD should probably be
undertaken with rigid, transpedicular instrumentation, particularly in
the revision situation (85).

There are two common approaches to anterior lumbar
interbody fusion (ALIF). The first, the open retroperitoneal approach,
employs a 5–10 cm incision to directly access the anterior spine. A
complete discectomy may then be undertaken and a variety of implants,
including allograft rings and threaded cages, placed into the disc
space. Various endoscopic methods of threaded cage placement have been
described as well. These approaches use relatively straightforward
techniques for access, but they do not include a complete discectomy.
Moreover, the threaded cage techniques require violation of the
vertebral endplate. While these newer approaches may be less invasive,
long-term data regarding fusion rates and implant stability are not
available.

Historically, ALIF has been reserved as a salvage procedure for patients failing multiple posterior procedures (27).
More recently, increased ease of access and concerns over extensor
muscle retraction in a relatively young patient population have renewed
interest in this approach. Moreover, some authors, citing the disc as
the primary source of pain, recommend its complete extirpation (27,29,43,58,59,72,82,89,90).

The advantages of ALIF include the following:

Advocates of ALIF report that even after solid posterior
arthrodesis, flexion may occur through the fusion mass. This slight
movement may cause continuing pain in the intervening degenerated
discs. Weatherly et al. (81) reported complete
relief of pain after ALIF in four patients with solid posterolateral
fusion and concordant discograms at the previously fused level. Because
ALIF places the fusion mass at the center of motion of the spine, it
more rigidly immobilizes the spine once it is solid (33,90).

Results of ALIF have included highly variable fusion rates from 18% to 96% (26).
However, the pseudarthrosis rate is generally felt to be lower than
that for one-level posterolateral fusions or posterior lumbar interbody
fusions (76,85).
Similar variability in rates of pain relief and return to work have
been reported. Ostensibly, this variability is caused by differences in
surgical indications and techniques.

Several distinct operative techniques have been described for ALIF (see Chapter 146).
Traditionally, corticocancellous bone from the iliac crest was placed
in the disc space and maintained in position with a screw and washer.
This approach was sometimes associated with graft collapse and
pseudarthrosis. Allograft rings were then recommended and,
subsequently, threaded cage techniques.

The cited advantages of threaded cages include
structural support for the anterior column, indirect decompression of
the foramina and nerve roots by distraction of the disc space, the
potential for bony ingrowth through the cage, and the possibility of
minimally invasive implantation (90). While clinical results are preliminary, early series demonstrate good mechanical stability with these constructs (52). Figure 145.15 shows sample cases.

When considering the more recently described endoscopic
methods of ALIF, it is important to remember that they represent only a
new technique, not a new operation. Therefore, operative indications
remain the same. There are proposed benefits of an endoscopic approach,
however. Preliminary studies suggest shorter hospital stays, decreased
morbidity, and earlier return to work with minimally invasive
techniques (52,53).

At present, several endoscopic techniques are evolving.
The transperitoneal approach with gas insufflation serves as a direct
extension of conventional laparoscopic surgery (90).
This technique allows direct access to L5–S1, L4–L5, and occasionally
L3–L4. Proposed advantages include ease of organ retraction, more rapid
exposure of the spine, increased working space, and decreased bleeding.
Disadvantages include the requirement for expensive trocars with
diaphragms and other special instruments, as well as the potential for
air leakage and a carbon dioxide venous embolism.

A gasless approach has been devised wherein the working
space is maintained by lifting the anterior abdominal wall with a
fan-like retractor. This approach allows the use of conventional
instruments and avoids carbon dioxide effects and expensive valves. The
procedure is associated with increased time for exposure, limited
lateral vision, and an overall more technically difficult approach.
Therefore, an intermediate, combined insufflated and gasless technique
was devised, in which insufflation is used for the initial spine
exposure. Retractors and Steinmann pins are placed, and then conversion
to gasless technique is undertaken.

For a retroperitoneal approach utilizing the potential
space between the spine and the abdominal cavity, make a 2.5 cm flank
incision by splitting the anterior lateral abdominal muscles. After an
initial finger dissection, enlarge the space with a retroperitoneal
balloon. Again, the laparoscopic retractor with a hydraulic arm is
employed to create a tent-like effect. Further peritoneal reflection
can then be carried out under direct vision. This procedure provides
access from T-12 to S-1. Further, conventional instruments may be used
and the procedure can be converted from a pure percutaneous endoscopic
to an endoscopic-assisted anterior approach should the degree of
difficulty be increased. Some authors report difficulty obtaining a
direct frontal approach to the disc space with this technique, however.

Global (360°) fusions were previously recommended for multilevel involvement and postlaminectomy patients. Kozack and O’Brien (44)
reported a series of 69 patients with global fusions for DDD. Fusion
levels were determined by provocative discograms. With one- and
two-level procedures, 90% fusion rates were achieved. Three-level and
greater procedures were associated with 77% fusion rates. Overall, 80%
had acceptable clinical results. O’Brien et al. (59)
described a global fusion procedure for 150 patients with severe
disability due to back pain or with previous failed operations. With
posterior instrumentation, they noted an 86% success rate.

Lower pseudarthrosis rates, usually 5% to 10%, were the
principal justification for the added surgical morbidity of these
combined procedures (44,76).
More recently, concerns regarding the effectiveness of threaded cages
as stand-alone devices has intensified the debate over the need for
posterior stabilization.

A variation on this theme involves an endoscopic anterior fusion with a percutaneous posterior stabilization

procedure with pedicle screws or translaminar facet screws, but there are no reports of long-term results.

A method of achieving an anterior arthrodesis with
posterior stabilization in a single surgical approach is the posterior
lumbar interbody fusion (PLIF). A wide posterior decompression is
performed, allowing retraction of the dural sleeve and nerve roots for
complete disc excision and anterior column fusion.

The PLIF was introduced by Cloward in 1945 to treat lumbar disc ruptures (9).
He stated that PLIF was indicated for “the treatment of low back pain
with or without sciatica due to lumbar disc disease.” This procedure
has also been recommended for spondylolisthesis, lumbar scoliosis,
osteomyelitis, lumbar kyphosis, and to increase posterior fusion rates
in high-risk patients (e.g., smokers and diabetics).

In each case, the addition of an anterior load-sharing
graft may enhance the fusion rate, stabilize the construct, and protect
the posterior spinal implant by load sharing. In patients with
deformity, the PLIF may aid in correction by partial anterior release.
Most commonly, however, PLIF procedures are performed for discogenic
back pain.

Cloward (9) summarized the PLIF
as able to address “all sources of pathologic change of the motion
segment in one operation, through one incision.” Yet, after initial
enthusiasm, use of the PLIF declined due to high rates of
pseudarthrosis and graft dislodgement. More recently, the advent of
transpedicular instrumentation led to a resurgence of interest in PLIF.
Steffee and Sitkowski (74) reported 104 fusions performed without graft dislocation, pseudarthrosis, or infection.

Reported advantages of PLIF include near total disc
excision, restoration of disc height and normal sagittal contour, root
decompression, solid mechanical arthrodesis, immediate load-sharing
with structural support, large surface area for fusion between the
vertebral endplates, fusion under compression, and avoidance of an
additional anterior approach.

Disadvantages include graft displacement,
pseudarthrosis, anterior and posterior destabilization, increased
bleeding, dural tears, risk to nerve roots, and risk of epidural
fibrosis from root retraction (36).

Ma (50) reported 100 consecutive
PLIFs for back or leg pain, spondylolisthesis, or failed back syndrome.
There was an 11% reoperation rate: six for pseudarthrosis, three for
bone graft extrusion, one bone graft fracture, and one hematoma. Others
have reported good to excellent results in 89%, with a fusion rates of
73% to 95% and an 82% return-to-work rate (49,76).

Of particular concern is the trauma to the nerve roots
from wide retraction. Epidural fibrosis may develop into a chronic
radiculopathy for which there is presently no satisfactory solution.
Harms et al. (33) recently popularized a
variant of the PLIF procedure that avoids significant retraction on the
thecal sac. This posterolateral approach to the disc space relies on
facet excision and has been termed the transforaminal lumbar interbody
fusion (TLIF). While transpedicular instrumentation is recommended for
the midline laminotomy version of the PLIF, it is mandatory here.

Transforaminal lumbar interbody fusion preserves the
anterior and posterior ligamentous complex, thereby maintaining a
tension band for compression of the graft and prevention of
retropulsion. While long-term outcome reports are not yet available,
some authors noted the proximity of the dorsal root ganglion and the
potential for chronic, neurogenic pain after even minor trauma to this
structure.

Posterior lumbar interbody fusion is contraindicated in
patients with preexisting, significant epidural fibrosis and those with
significant osteopenia. Aside from the more typical complications such
as bleeding, infection, and pseudarthrosis (mentioned later), the
unique complications possible with PLIF deserve special mention here.
With standard PLIF procedures, damage to nerve roots remains a
principal concern. The surgeon must be careful about overdistraction,
particularly in patients with nerve root anomalies. New or increased
deficits occur in 0.5% to 4% of patients after PLIF (48).
The upper (exiting) root traverses the interspace just out of direct
view in the lateral recess and can be damaged when grafts are inserted
in the disc space.

Steffee and Sitkowski (74) report that a failed PLIF has

“a worse outcome than failure of any other fusion procedure.” They
report that exploration of patients with a post-PLIF chronic
radiculopathy reveals epidural fibrosis for which there is no
satisfactory salvage. Careful attention to the tension placed on the
cauda and nerve roots may diminish the incidence of these problems.

Neural structures may also be injured with posterior
bone graft migration. The subtotal discectomy required for a PLIF
procedure risks penetration of the anterior annulus with attendant
anterior vessel damage, a potentially catastrophic complication.

Revision options in failed PLIF are limited. A
noninstrumented PLIF may be converted to an instrumented PLIF. A failed
instrumented PLIF is most often revised with an attempted anterior
fusion (82).

With the rising popularity of anterior approaches in the
treatment of lumbar disc disease, the role of PLIF or TLIF is not
entirely clear. For example, the importance of “fusion disease” in
these relatively young patients remains to be established (42). Certain criteria, however, may reasonably aid in the selection between ALIF and PLIF (Table 145.9).

Because PLIF is a relatively difficult surgery to
revise, present recommendations for PLIF include lumbar disc disease
with sciatica (10) and certain revision situations in which an anterior/posterior fusion through a single incision is desirable.

We feel that a posterior fusion with instrumentation is
the gold standard for the rare patient with true lumbar segmental
instability. Those patients with painful lumbar spondylosis LS are best
served with a disc ablative procedure such as ALIF and PLIF.
Circumferential fusions should be reserved for patients with
significant canal pathologies, where revision is required, and other
situations in which additional stabilization is required.

We favor ALIF because of the destabilizing effects of
PLIF and its potential for nerve root injury, epidural fibrosis, and
“fusion disease.” The advent of transforaminal interbody fusion
procedures may decrease the risk of the PLIF technique, but firm
evidence is not available. As more information regarding the long-term
results of interbody cage placement becomes available, treatment
recommendations will no doubt be revised.

Before undertaking a posterior spinal fusion for DDD,
carefully review your patient selection criteria. Be sure the patient
is well informed as to the risks and the potential for an
unsatisfactory clinical outcome. Pay careful attention to patient
physiology by the following measures:

A posterolateral fusion begins with proper patient
positioning. Choose a frame or bolsters that allow the abdominal
contents to be free of compression and maintain proper lumbar lordosis (Fig. 145.16).

As in all spine surgery, a headlight and loupes increase
the effectiveness of the exposure. Intraoperative neural monitoring may
be used to stimulate pedicle screws during placement to detect
pedicular penetration.

Spine fusion procedures may be associated with
significant blood loss. Intraoperative blood salvage is occasionally
useful. More important, anticipate and control bleeder sites as you
encounter them. Three fairly constant bleeding points include the pars
interarticularis artery, the artery of Macnab (transverse process
artery), and the sacral arteries.

If no midline decompression is required, consider a
bilateral Wiltse (paramedian) approach. A larger midline skin incision
and subcutaneous dissection is required. Make the fascial incisions two
finger breadths lateral to the midline bilaterally. Then bluntly
dissect down to the lateral facets between the multifidus and
longissimus muscles. This approach is covered more extensively in Chapter 138.

Employ the same initial steps as in a noninstrumented
fusion. Do not insert pedicle screws prior to complete exposure of the
posterior elements. Decortication of the transverse processes and bone
graft insertion are more easily and completely accomplished prior to
hardware insertion. Many surgeons, however, prefer to decorticate after
instrumentation to diminish blood loss.

Several useful techniques for entry-point localization
and pedicle-screw placement have been described. Thorough knowledge of
pedicular anatomy is critical. Various radiographic techniques of
localization are now available, but none replaces a firm grasp of the
relationships of the posterior elements to one another.

The pedicle lies along the midline axis of the
transverse process. The outer border of the facets describes a line
roughly conforming to the lateral border of the pedicle. The lateral
border of the pars defines the medial pedicular border. The accessory
process is often identified at the junction of the facet and the
transverse process. This process serves as a useful entry point (Fig. 145.20A).

While multilevel fusions are generally not recommended
for DDD, instrumentation in these cases is placed at each level in an
effort to reduce micromotion. How much segmental stiffness is required
to achieve fusion and eliminate pain is not established (45); however, given that

micromotion has been hypothesized as a cause of pain in four screw micromotion-segment constructs (81),
maximizing points of fixation is recommended. Screw placement is
abandoned if preoperative imaging demonstrates thin pedicles.
Similarly, if difficulties with cortical breech are noted
intraoperatively, abandonment of that point of fixation is recommended.

Sacral fixation remains the weak link in posterior fixation systems (13).
Large constructs extending to the sacrum require additional fixation.
The combination of large forces transmitted to the sacrum and the size
and bone content of the sacral pedicle may render S-1 screws inadequate
as the sole point of inferior fixation. In the degenerative conditions
described in this chapter, however, smaller, single-level constructs
are recommended. In these cases, simple transpedicular instrumentation
is usually successful. The sacral pedicle requires larger screws.

Medial angulation allows longer screws with superior
fixation strength, although iliac crest overhang may limit optimal
screw trajectory. In the case of tenuous fixation, additional cortical
purchase is recommended to prevent “windshield-wipering.” Options
include careful perforation of the anterior cortex or of the L5–S1 disc
space (Fig. 145.21).

When considering penetration of the anterior sacral cortex, recognize the possibility of significant individual

anatomic variability (13).
Differences in sacral bony anatomy are associated with differences in
vascular tree branching patterns and in the location of the neural
foramina. Assess the morphology and choose an optimal trajectory from
preoperative CT or MRI scans.

The medial sacral promontory is the safest place to
penetrate the anterior cortex. More lateral sacral screws may hit the
lumbosacral plexus, which is affixed to the bone here. Lateral
trajectories also risk injury to the iliac veins, particularly on the
left.

Other options for fixation below the lumbosacral
interspace include S-2 pedicle screws, sacroiliac bolts, and sacral
rods. S-2 screws are easily inserted along the intermediate sacral
crest midway between the first and second dorsal foramina, but they are
considered biomechanically weak. Cortical penetration at this level
risks injury to the sigmoid colon.

Postoperative bracing is not usually recommended, and
genuinely effective bracing requires a pantaloon extension. Sacral
fixation techniques are further discussed in Chapter 156, Chapter 159 and Chapter 160.

Carry out the initial stages of PLIF (10,48,74)
in the same manner as posterolateral fusion. Careful positioning is
critical. Lumbar lordosis and abdominal decompression are important.
Some authors feel the knee–chest position adequately maintains lordosis
while maximizing hip and

knee
flexion. Hip and knee flexion ensure decreased tension on the nerve
roots, which may allow greater thecal sac retraction for PLIF surgery.

Although PLIF was originally described without
instrumentation, this is not recommended. The wide posterior exposure
necessary for a safe PLIF produces increased instability with increased
rates of pseudarthrosis and graft dislodgement.

Classically, a PLIF is performed in three steps: laminotomy, removal of the intervertebral disc, and spinal fusion.

A number of techniques are described for preparing the
disc space. Specialized instruments are particularly useful. If they
are not available, curets, pituitary rongeurs, and osteotomes may be
used. The goal is to clear 80% to 90% of the disc space. Better
preparation of the graft site yields a larger area of bony contact with
the graft and increases the chance of successful fusion.

Numerous techniques are also described for the posterior grafting of the anterior column (Fig. 145.23).
With disc space distraction maintained, various graft materials may be
inserted. These include dowels, corticocancellous ramps, titanium
plugs, and metal or carbon fiber cages of the vertical (e.g., the Harms
cage, DePuy-Acromed, Raynham, MA) or the horizontal variety (e.g., the
BAK, Sulzer-Spinetech Minneapolis, MN) (12).

Cages are designed to prevent postoperative collapse by
providing a structural support to the interspace while the cancellous
graft material becomes incorporated. The use of morcelized cancellous
bone rather than corticocancellous pieces may minimize donor site
morbidity. Specialized instrumentation has been developed for the
posterior insertion of threaded interbody cages. While the specific
instruments vary by the system used, the concepts are the same. The
larger size of these implants requires a larger laminotomy or complete
laminectomy and facetectomy. Base implant sizing on preoperative
templating. Furthermore, the larger size of the instrumentation
warrants a heightened awareness of the position of and tension on the
dura and nerve roots.

The TLIF variant (33) of the standard PLIF procedure requires a wide, unilateral approach to the disc space at its posterolateral aspect.

Regardless of the approach, early mobility is a prime postoperative goal.

Most patients are ready for discharge by the third
postoperative day. See them for a wound check and staple removal at
7–14 days after surgery. Subsequent visits include 6- and 12-week
checks. Obtained radiographs at these intervals to assess the status of
the fusion.

The schedule for return to work and sporting activities
is individualized on the basis of clinical status, radiographic union,
and the nature of the activities involved. Early return with limited
duty is preferred to lengthy sick leave.

After solid fusion is achieved, aerobic activities,
spinal flexibility, and truncal strength and stability are important,
life-long aspects of the patient’s personal fitness regime. Further
specific physical therapy is not usually needed, however.

Some patients may not be able to return to physically
demanding occupations. Early intervention with skills retraining and a
realistic outlook are crucial to optimal functional recovery.