Abstract
Background: The purpose of the present study
was to determine the effectiveness of selective nerve-root injections in
obviating the need for an operation in patients with lumbar radicular
pain who were otherwise considered to be operative candidates. Although selective
nerve-root injections are used widely, we are not aware of any prospective,
randomized, controlled, double-blind studies demonstrating their
efficacy.
Methods: Fifty-five patients who were referred
to four spine surgeons because of lumbar radicular pain and who
had radiographic confirmation of nerve-root compression were prospectively
randomized into the study. All of the patients had to have requested operative
intervention and had to be considered operative candidates by the
treating surgeon. They then were randomized and referred to a radiologist who
performed a selective nerve-root injection with either bupivacaine
alone or bupivacaine with betamethasone. The treating physicians
and the patients were blinded to the medication. The patients were
allowed to choose to receive as many as four injections. The treatment
was considered to have failed if the patient proceeded to have the
operation, which he or she could opt to do at any point in the study.
Results: Twenty-nine of the fifty-five patients,
all of whom had initially requested operative treatment, decided
not to have the operation during the follow-up period (range, thirteen
to twenty-eight months) after the nerve-root injections. Of the
twenty-seven patients who had received bupivacaine alone, nine elected
not to have the operation. Of the twenty-eight patients who had
received bupivacaine and betamethasone, twenty decided not to have
the operation. The difference in the operative rates between the
two groups was highly significant (p < 0.004).
Conclusions: Our data demonstrate that selective
nerve-root injections of corticosteroids are significantly more effective
than those of bupivacaine alone in obviating the need for a decompression
for up to thirteen to twenty-eight months following the injections
in operative candidates. This finding suggests that patients who
have lumbar radicular pain at one or two levels should be considered
for treatment with selective nerve-root injections of corticosteroids
prior to being considered for operative intervention.
Since Macnab10 described
the technique of selective nerve-root injection twenty-nine years
ago, numerous investigators have reported on its value in treating patients
with radicular pain9,12-14. However,
there has not been, as far as we know, a single prospective, randomized,
double-blind, controlled study evaluating its efficacy for the treatment
of lumbar radicular pain.
Most authorities agree that the initial treatment of acute radicular
pain in the low back should be bed rest, anti-inflammatory medication,
and physical therapy7,10,14. However,
many patients who have been so treated have persistent pain and
seek further intervention. An epidural injection of steroids is
a popular treatment method, despite the fact that the short and
long-term results of such therapy remain controversial3,5,7,14. An alternative method for
delivering steroids to inflamed nerve tissue is the use of selective
nerve-root injection, which is the term given to the procedure developed
by Krempen and Smith9 twenty-six
years ago. Under fluoroscopic guidance, the needle is placed next
to the presumed affected nerve root, resulting in a precise and concentrated
delivery of the drug to the immediate vicinity of that nerve. While
nerve-root injections have been demonstrated to diminish lumbar
radicular pain in a number of nonrandomized studies9,12-14, one could ask whether some
of the results might not have been attained with analgesic treatment alone.
A more important question is whether these injections can obviate
the need for an operation in patients who would otherwise be operative
candidates. Although it is effective, operative treatment of lumbar
radicular pain is associated with both morbidity and mortality.
In addition, the overall cost of operative treatment of lumbar radicular pain,
including the cost of the operation and lost wages, is exceedingly
high. Because of the relatively low cost and morbidity associated
with the injections compared with those associated with an operative
procedure, injections are likely to be cost-effective if they allow
even a small percentage of patients with lumbar radiculopathy to
avoid operative intervention.
The purpose of the present study was to critically examine the
efficacy of selective nerve-root injections of corticosteroids in
a prospective, randomized, controlled, double-blind manner to determine
if they could obviate the need for operative intervention in patients
with lumbar radiculopathy who otherwise would be operative candidates.
The null hypothesis was that there is no significant difference
between patients treated with bupivacaine alone and those treated
with bupivacaine and steroids with respect to the eventual operative
rate.
Inclusion and Exclusion Criteria
The study subjects were selected from patients who were seen
at the offices of four spine surgeons (three orthopaedic surgeons
and one neurosurgeon) at one institution during a one-year period from
December 1996 to December 1997. All patients who were more than
twenty-one years old and who had degenerative lumbar radicular pain with
a disc herniation or central or foraminal spinal stenosis confirmed
by magnetic resonance imaging or computed tomography-myelography were
eligible. There were no differences between the experimental and
control groups with regard to any of the measured variables, including
age, gender, number of levels of disease, diagnosis (herniated nucleus
pulposus or stenosis), and number of previous operations. In addition,
we included patients who had had a previous operation and demonstrated
persistent or new neurological compression. They had to have completed
a course of nonoperative management including anti-inflammatory
medication, physical therapy, and activity modification for at least
six weeks without adequate benefit. An exception to the six-week rule
was made only if the patient had intractable pain despite the use
of the maximum dosage of anti-inflammatory medication, supplemented
with narcotic pain medication, and activity modification. Both the
patient and the treating surgeon had to agree that operative treatment
was the next indicated step. Prior to the era of the use of nerve-root injections,
operations would have been scheduled for all of these patients.
The criteria for exclusion were acute trauma, cauda equina syndrome,
a progressive neurological deficit, a motor deficit, a pathological
or infectious etiology, a patient who was not an operative candidate,
involvement with a Workers' Compensation claim, a history of an
adverse reaction to corticosteroids or local anesthetics, lack of
a radiographically detectable abnormality, more than two radiographically
abnormal and symptomatic levels on either side such that three or
more separate injections would be necessary to alleviate the symptoms,
and an absence of substantial radicular pain as the presenting symptom.
All patients who met the criteria were offered enrollment in
the present study, and a total of fifty-five patients agreed to
participate. There were twenty-seven men and twenty-eight women.
After signing an informed consent approved by the Human Studies
Committee at our institution, the enrolled patients had a thorough
initial physical examination by the spine surgeon. Then they filled out
a nerve-root-injection questionnaire, and a North American Spine
Society low-back-pain outcome questionnaire6 was
administered by a nurse. On the basis of the patient's symptoms,
signs, and radiographic findings, the surgeon picked one or, at
most, two levels to be injected. The patients were randomly assigned
to the experimental or the control group by the study administrator
(Y. Y.), a full-time radiology research fellow with no involvement
in patient care. They were then referred to our radiology department
for a selective nerve-root injection with bupivacaine alone (control
group) or bupivacaine with betamethasone (experimental group). One
of three radiologists, each experienced in the injection technique,
performed all of the injections. Care was taken to try to ensure
that all of the injections were performed in as identical a manner
as possible. All of the injections were performed according to a
strict protocol (described below) and after a contrast radiculogram
had been made to ensure that the correct nerve root was being injected. Neither
the patient nor the treating spine surgeon knew the group to which
the patient had been randomized. Just prior to the procedure, each
patient completed a nerve-root-injection questionnaire indicating
his or her pain level on a scale of 0 to 100 points (with 100 points
indicating the worst pain and 0 points, no pain).
Injections
The patient was placed prone on a fluoroscopic table with c-arm
capability to allow adjustment of the angle of the x-ray beam in
order to visualize the target area in posteroanterior, oblique,
and lateral positions. An entry site that allowed visualization
of the lateral edge of the pars interarticularis, transverse process,
and articular facets on the side to be injected was selected within
a few centimeters lateral to the spine. Usually, the disc space at
that level was also profiled. The entry site was marked with indelible
ink. Sterile preparation was performed with alcohol followed by
Betadine (povidone-iodine) solution. Local superficial anesthesia
with 1 percent lidocaine buffered with a small amount of bicarbonate
was injected. A 22-gauge, five-inch (thirteen-centimeter) spinal
needle with a short bevel, which had its tip bent to allow easier
entry into the foramen, was placed, under fluoroscopic control,
into the anterosuperior portion of the selected lumbar foramen.
Contrast solution was injected, under fluoroscopic control, through
a connecting tube to verify that contrast material and subsequent
medication were passing about the desired nerve root and also to
be sure that the injectant was not flowing intravascularly. The connecting
tube then was removed, and the selected injectant was delivered
with a single syringe. Radiculogram radiographs for final verification were
made for the permanent records, and a report was generated, but
the contents of the injectant were not noted in the report.
Prior to the procedure, the results of the randomized selection
had been given by the project coordinator to the radiologist. All
patients received one milliliter of 0.25 percent bupivacaine, while
those who had been selected for treatment with steroids also received
one milliliter of betamethasone (six milligrams per milliliter).
In either case, the injectant was delivered by means of one syringe. Following
the procedure, the needle was removed, the site was cleaned, and
Betadine (povidone-iodine) ointment and a bandage were applied.
The patient then was taken from the radiology suite, and a form
for follow-up measurement of pain and an instruction sheet describing
the procedure were given to the patient to take home.
Failure of Treatment
The patient was allowed to choose to receive as many as four
injections at any time during the follow-up period. This is consistent
with our normal routine for treating patients with corticosteroid
injections. If one or more additional injections were requested,
the patient received the same medication that he or she had received
for the first injection, in a double-blind manner. The injection therapy
was not considered to have failed if the patient received two or
three additional injections; however, it was considered to have
failed if the patient opted for operative treatment.
Data
The treating surgeon was blinded to the medication throughout
the entire period of the study. The orthopaedic nurse conducted
a complication survey two weeks after the injection. The first follow-up
evaluation by the surgeon was at one to four weeks following the
injection. The follow-up questionnaires, including the nerve-root-injection and
North American Spine Society questionnaires, were mailed to the
patient eight weeks after the injection. One year after the injection,
a follow-up evaluation was performed in person or by telephone,
and the two questionnaires were administered again by a nurse. Final
scores were calculated separately for each section of the North American
Spine Society questionnaire. In February 1999, final follow-up telephone
calls were made to all of the patients who had not, to our knowledge,
had an operation, to make sure that they had not had an operation
performed by a different surgeon. The control and experimental groups
were compared with use of chi-square tests. The findings before
the procedure and the result at one year after the injection were
compared with use of a paired t test.
Measurement of Outcome Data with Use of the North
American Spine Society Outcome Instrument
The North American Spine Society outcome instrument included
questions regarding medical history, expectations, and outcome6. The follow-up questionnaire included
questions on whether the patient's baseline expectations had been
met following treatment.
We compared the baseline and follow-up questionnaires for all
of the patients who had avoided an operation. We analyzed these
data in two ways. First, we calculated the mean scores at baseline and
compared them with the mean scores at the final follow-up evaluation.
We then performed a t test to determine if there was a significant
difference. In addition, we used the recommendations of the North
American Spine Society outcome measurement tool6,
which states that a difference of 1 point in any subsection is significant.
We counted the number of individuals in each group for whom the
score had increased or decreased by 1 point.
We stratified the data further to compare the baseline with the
follow-up data for patients with a diagnosis of stenosis as well
as for those with a diagnosis of herniated nucleus pulposus. We
assumed that the injections might help the patients with a herniated
disc more than those with stenosis.
Comparison of Baseline Data Between Experimental
and Control Groups
Twenty-eight patients were injected with bupivacaine and betamethasone,
and twenty-seven were injected with bupivacaine only. No differences
between the experimental and control groups were detected with regard
to any of the measured variables, including age, gender, number
of levels of disease, diagnosis (herniated nucleus pulposus or stenosis),
and number of previous operations. In addition, no significant differences
between the experimental and control groups were found with respect
to the baseline North American Spine Society outcome measurements.
Follow-up
A final follow-up evaluation was performed for all fifty-five
patients at thirteen to twenty-eight months (mean, twenty-three
months) following the first injection. The patients who had opted
not to have an operation were followed for a minimum of fifteen
months (mean, twenty-three months; range, fifteen to twenty-eight
months). As a group, the patients with stenosis who had opted not
to have operative treatment had the longest duration of follow-up
(range, sixteen to twenty-eight months). The duration of follow-up
for patients with a herniated disc ranged from fifteen to twenty-eight
months. The patients who had an operation were followed for a mean
of twenty-two months (range, thirteen to twenty-eight months).
All of the patients who succeeded in avoiding operative treatment
filled out the North American Spine Society low-back-pain outcome
questionnaire at a minimum of one year following the injection.
Operative Treatment (Failure of Injection Treatment)
Twenty-six of the fifty-five patients opted to proceed with an
operation, and twenty-nine chose not to have one. Of the twenty-seven
patients who had received bupivacaine only, eighteen opted for operative
intervention whereas nine did not. Of the twenty-eight patients
who were treated with bupivacaine and betamethasone, only eight
proceeded with an operation while twenty chose not to have one.
The difference was highly significant (p < 0.004). Stated another
way, of the twenty-six patients who opted for operative intervention,
eighteen had received bupivacaine alone and eight had received bupivacaine
and betamethasone.
Comparison of Baseline and Follow-up Questionnaires
for All Patients Who Avoided Operative Treatment
Although there was a trend toward a decrease in the neurological
symptoms from the baseline to the final follow-up evaluation in
all of the patients who avoided operative treatment, the difference was
not significant. When we analyzed the data in the subgroups of patients
who had stenosis and those who had herniated nucleus pulposus, we found
that the patients with stenosis had a significant decrease in neurological
symptoms from the baseline to the final follow-up evaluation (p < 0.03).
We also found that the patients with stenosis had significant relief
of low-back pain between the baseline and final follow-up evaluations
(p < 0.008) and that the patients with herniated nucleus pulposus
had a trend toward relief (p < 0.07). The most dramatic improvement
was reflected by the patients' responses regarding their treatment
expectations, with significant improvement from the baseline to
the final follow-up evaluation in both the group that had stenosis
(p < 0.00004) and the group that had herniated nucleus pulposus
(p < 0.000018). The job-exertion and job-stress values were
unchanged.
Nine patients who received bupivacaine only continued with nonoperative
treatment. The only significant difference between the baseline
and follow-up outcome data in this subgroup was in their scores
on the questions regarding treatment expectations (p < 0.001).
The patients who had a diagnosis of stenosis and were injected with
bupivacaine and betamethasone had significant relief of the low-back
pain (p < 0.049) and significant improvement in their scores
on the questions regarding treatment expectations (p < 0.002).
The patients who had herniated nucleus pulposus also had a significant
difference, between the baseline and follow-up evaluations, in their
scores on the questions regarding treatment expectations (p = 0.001).
Multiple Injections
Nineteen patients received more than one injection: ten had two
injections, six had three injections, and three had four injections.
The minimum interval between the injections was six days; the maximum
was 10.5 months. Thirteen of the nineteen patients who had multiple
injections avoided operative treatment, while six did not. The difference
almost reached significance (p = 0.059).
We found that nerve-root injections were indeed effective in
obviating the need for an operation in more than half of the operative
candidates. Twenty-nine of the fifty-five patients who had requested operative
intervention and who were considered to be operative candidates
by their treating spine surgeons avoided an operation. Compared
with injection of bupivacaine alone, injection of bupivacaine with
the steroids was significantly (p < 0.004) more likely to result
in the avoidance of an operation. Interestingly, nine of the twenty-seven
patients who had received bupivacaine alone still managed to avoid
an operation.
We were surprised that so many of our patients who received bupivacaine
with steroids avoided an operation. Normally, all of these patients
would have been treated operatively. Unfortunately, our data do
not give us a good explanation as to why. We propose several possible
theories. First, it may be that any new treatment modality that
is added to the existing treatment regimen may help to decrease
the need for operative intervention; a similar study comparing the
results of any new interventional treatment may have yielded similar results.
Second, most patients who have sciatica do not have steady, intractable
pain but, rather, intermittent, varying degrees of pain. Patients
may have intolerable pain for just a few hours during the day or
perhaps for just a few days during the week. We found that the patients
who proceeded to have an operation had higher baseline scores for back
pain and neck pain than did those who avoided an operation, regardless
of the treatment group. Treatment with bupivacaine with or without
corticosteroids certainly can diminish the pain in the short term.
As long as the patient feels that there are alternatives to an operation
that can diminish the pain, no matter how transiently, they may
feel that they have enough control over the pain to avoid the more
aggressive route. This theory is supported by the fact that there
was a highly significant improvement in the scores for the questions
on "expectations of improvement" by the time of the final follow-up.
Third, even asymptomatic individuals can have radiographic evidence
of stenosis2,15, and patients
who are symptomatic may improve either spontaneously or with anti-inflammatory medications
and physical therapy. As stated previously, these are mere conjectures;
we do not really know why so many patients were able to avoid an operation.
In previous studies in which only the degree of pain relief was
measured1,4,8,11, several of our
patients who managed to avoid operative intervention would have
been considered to have had failure of treatment. Indeed, some of these
patients actually had a higher score for leg or back pain on the
final follow-up evaluation compared with that on the baseline evaluation.
The difference is that no patient had a higher (worse) score for
pain in more than one category, and every patient who avoided an
operation showed a highly significant improvement in the scores
on the expectation questions.
Limitations of the Study
As with any study, the present investigation may have several
potential problems. Because of the relatively small number of patients
who were randomized, any absence of statistical significance that
we found may be suspect. In other words, the absence of statistical
significance may be due to a true lack of difference or to an inability
to detect small differences because of an insufficient sample size,
or to both. On the other hand, if the sample size had been larger,
we may have been able to detect a difference between the patients
who had herniated nucleus pulposus and those who had spinal stenosis
with respect to the outcome measures. It is recognized that patients
with herniated nucleus pulposus can fully recover spontaneously.
In contrast, stenosis is expected to gradually worsen with time,
so the effects of injection treatment may eventually disappear.
Our follow-up period of fifteen to twenty-eight months for the patients
with stenosis may not be adequate to determine if these patients
will avoid an operation in the long term. While it is likely that
a patient with herniated nucleus pulposus who avoids operative treatment
for fifteen months has a reasonable chance of avoiding operative
treatment forever, it is difficult to say the same about a patient
who has spinal stenosis. Given the fact that patients who have spinal
stenosis tend to be older, postponing the operation until they are
even older and perhaps more medically unstable may not be in their
best interest. For this reason, we plan to follow these patients
during the next few years to determine whether they are able to
continue to avoid operative intervention.
As a direct result of the present study, our treatment algorithm
for patients who have one or two-level radiculopathy in the lumbar
spine without substantial motor or sensory deficits now routinely includes
the use of selective nerve-root injections of bupivacaine and corticosteroids.
Only if they refuse the injections or fail to improve after a maximum
of three or four such injections are they offered operative intervention.
Given the highly significant difference between the control and
the experimental group in this prospective, randomized, controlled,
double-blind study, we recommend that this algorithm be more widely considered.
Berman, A. T.; Garbarino, J. L., Jr.; Fisher, S. M.; and Bosacco, S. J.: The effects of epidural injection of local anesthetics
and corticosteroids on patients with lumbosciatic pain. Clin. Orthop.,,188: 144-151, 1984.188144
1984
Boden, S. D.; Davis, D. O.; Dina, T. S.; Patronas, N. J.; and Wiesel, S. W.: Abnormal magnetic-resonance scans of the lumbar spine
in asymptomatic subjects. A prospective investigation. J. Bone and Joint Surg.,72-A: 403-408, March 1990.72-A403
1990
Bogduk, N.: Epidural steroids. Spine,,20: 845-848, 1995.20845
1995
Bush, K., and Hillier, S.: A controlled study of caudal epidural injections of triamcinolone
plus procaine for the management of intractable sciatica. Spine,,16: 572-575, 1991.16572
1991
Cuckler, J. M.; Bernini, P. A.; Wiesel, S. W.; Booth, R. E., Jr.; Rothman, R. H.; and Pickens, G. T.: The use of epidural steroids in the treatment of lumbar
radicular pain. A prospective, randomized, double-blind study. J. Bone and Joint Surg.,67-A: 63-66, Jan 1985.67-A63
1985
Daltroy, L. H.; Cats-Baril, W. L.; Katz, J. N.; Fossel, A. H.; and Liang, M. H.: The North American Spine Society lumbar spine outcome
assessment instrument: reliability and validity tests. Spine,21: 741-749, 1996.21741
1996
[PubMed]
Derby, R.; Kine, G.; Saal,
J. A.; Reynolds, J.; Goldthwaite, N.; White, A. H.; Hsu, K.;, and Zucherman, J.:: Response to steroid and duration of radicular pain as
predictors of surgical outcome. Spine,,17 (Supplement 6): 176-S183, 1992.17 (Supplement 6)176
1992
Koes, B. W.; Scholten, R. J.; Mens, J. M.; and Bouter, L. M.: Efficacy of epidural steroid injections for low-back pain
and sciatica: a systemic review of randomized clinical trials. Pain,63: 279-288, 1995.63279
1995
[PubMed]
Krempen, J. F., and Smith, B. S.: Nerve-root injection. A method for evaluating the etiology
of sciatica. J. Bone and Joint Surg.,56-A: 1435-1444, Oct 1974.56-A1435
1974
Macnab, I.: Negative disc exploration. An analysis of the causes of
nerve root involvement in sixty-eight patients. J. Bone and Joint Surg.,53-A: 891-903, July 1971.53-A891
1971
Mathews, J. A.; Mills, S. B.; Jenkins, V. M.; Grimes, S. M.; Morkel, M. J.; Mathews, W.; Scott, C. M.; and Sittampalam, Y.: Back pain and sciatica: controlled trials of manipulation,
traction, sclerosant and epidural injections. British J. Rheumatol.,26: 416-423, 1987.26416
1987
Stanley, D.; McLaren, M. I.; Euinton, H. A.; and Getty, C. J.: A prospective study of nerve root infiltration in the diagnosis
of sciatica. A comparison with radiculography, computed tomography,
and operative findings. Spine,15: 540-543, 1990.15540
1990
[PubMed]
Tajima, T.; Furukawa, K.; and Kuramochi, E.: Selective lumbosacral radiculography and block. Spine,5: 68-77, 1980.568
1980
[PubMed]
van Tulder, M. W.; Koes, B. W.; and Bouter, L. M.: Conservative treatment of acute and chronic nonspecific
low back pain. A systematic review of randomized controlled trials
of the most common interventions. Spine,,22: 2128-2156, 1997.222128
1997
Wiesel, S. W.; Tsourmas, N.; Fetter, H. L.; Citrin, C. M.; and Patronas, N.: A study of computer-assisted tomography. I. The incidence
of positive CAT scans in an asymptomatic group of patients. Spine,,9: 549-551, 1984.9549
1984