This copy is for personal use only. To order printed copies, contact reprints@rsna.

org

Reviews and Commentary  n  How

Spinal Injections for Pain
Management1

I Do It
William E. Palmer, MD
Image-guided spinal injection is commonly performed in
symptomatic patients to decrease pain severity, confirm
the pain generator, and delay or avoid surgery. This arti-
cle focuses on the radiologist as spine interventionist and
addresses the following four topics relevant to the radiolo-
gist who performs corticosteroid injections for pain man-
agement: (a) the rationale behind corticosteroid injection,
(b) the interaction with patients, (c) the role of imaging
in procedural selection and planning, and (d) the pearls
and pitfalls of fluoroscopically guided injections. Factors
that contribute to the success of a pain management ser-
vice include communication skills and risk mitigation. A
critical factor is the correlation of clinical symptoms with
magnetic resonance (MR) imaging findings. Radiologists
can leverage their training in MR image interpretation to
distinguish active pain generators in the spine from inci-
dental abnormalities. Knowledge of fluoroscopic anatomy
and patterns of contrast material flow guide the planning
and execution of safe and effective needle placement.

q
 RSNA, 2016

Online supplemental material is available for this article.

1
 From the Department of Musculoskeletal Radiology,
Massachusetts General Hospital, 55 Fruit St, YAW 6030,
Boston, MA 02114. Received September 17, 2015;
revision requested October 23; revision received January
26, 2016; accepted February 19; final version accepted
March 30. Address correspondence to the author (e-mail:
wpalmer@mgh.harvard.edu).

q
 RSNA, 2016

Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 669

HOW I DO IT: Spinal Injections for Pain Management Palmer
I
n 1930, epidural anesthetic injec-
tion was described in the treatment
of sciatica (1). Epidural steroid
injection (ESI) was first performed
in the 1950s, it evolved as a thera-
peutic option during the 1960s, and
it became a cornerstone in the man-
agement of low back pain and sciatica
in the 1970s (2–6). During these de-
cades, needle placement and injection
site depended on palpated landmarks
and loss-of-resistance techniques. In
the 1980s and 1990s, radiologists and
Essentials
nn Structured interactions during
history taking, intervention, and
discharge build patient-physician
relationships that promote trust
and create rewarding opportu-
nities for radiologists to counsel
patients and affect clinical
decision making.
nn Correlation of symptoms and im-
aging findings guides targeted
inspection of MR images, differ-
entiation of active pain genera-
tors in the spine from painless
structural abnormalities, and for-
mulation of a treatment strategy.
nn Radiologists can leverage their
training in MR image interpreta-
tion, their knowledge of fluoro-
scopic anatomy, and their experi-
ence with contrast material
administration to plan and exe-
cute safe and effective needle
placement.
nn Therapeutic outcomes depend on
four factors: symptoms correlate
with MR imaging findings, symp-
toms result from inflammation,
inflammation is reversible, and
corticosteroid reaches the
inflamed tissue.
nn Corticosteroid injection has been
performed for decades for pain
management; however, it is con-
sidered an off-label use in spine
intervention in the United States
because the Federal Drug Admin-
istration has not approved corti-
costeroids for epidural or epira-
dicular administration.
670 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016
anesthesiologists began to use fluo-
roscopy to determine the accuracy of
caudal and interlaminar needle place-
ment, and they used epidurography to
understand patterns of injectate flow.
They found needle misplacements in
25%–38% of blind procedures per-
formed by experienced injectionists
(7–9). Epidurography was necessary
to confirm injectate location and to ex-
clude intravascular or intrathecal ad-
ministration (10–13).
Computed tomography (CT) and
magnetic resonance (MR) imaging
spurred growth in intervention by revo-
lutionizing the noninvasive diagnosis of
pain generators (14,15). When spinal
stenosis, disk herniation, and facet ar-
thropathy correlated with symptoms,
they were targeted for therapeutic in-
terventions. When the correlation was
uncertain, they were targeted for a sys-
tematic series of diagnostic interven-
tions. ESIs shifted from caudal routes
performed blindly to lumbar and cer-
vical routes directed at the imaging ab-
normalities and presumed pain sourc-
es (16). As fluoroscopic techniques
evolved, facet injection, nerve root
block (NRB), and discography were
added to lumbar and cervical ESIs (17).
During the 1980s and 1990s, uti-
lization data showed dramatic volume
growth as spinal interventions gained
widespread acceptance (18,19). Lum-
bar ESI rates for spinal stenosis in-
creased 300% within 2 decades (from
1994 to 2011) (20–23). By 2010, more
than 2.2 million lumbar ESIs were per-
formed yearly in Medicare patients
(21). Facet injections surged 147%
from 1993 to 1999 and increased an-
other 300% from 1998 to 2006 (23–
26). Medicare payments for spinal in-
jections expanded 629% from 1994 to
2001 (23). Before 2000, anesthesiolo-
gists performed the majority of injec-
tions (26). By 2007, procedures were
performed by anesthesiologists (49%),
physiatrists (25%), family practitioners
(12%), orthopedists (6%), and radiolo-
gists (3%) (27). Within the pain man-
agement domain, a small percentage of
providers performed a disproportion-
ately high percentage of spinal inter-
ventions (22,23,26,27).
Overuse led to the scrutiny of ther-
apeutic outcomes and the publication
of contradictory articles that defended
or criticized injections depending on
the interests of the authors (16,26,28–
38). Epidemiologists tended to focus
on long-term outcomes and surgical
end points. Injectionists focused on
short-term outcomes, accepting the
fact that interventions could modulate
but not cure the underlying cause of
symptoms. Research studies remain
difficult to compare due to disparate
symptoms (back pain vs radiculopathy),
diagnoses (spinal stenosis vs disk her-
niation), injection types (interlaminar
vs transforaminal), procedural tech-
niques (blind injection vs fluoroscopic
guidance), patient demographics, phar-
maceutical agents, and drug doses
(7,30,32–34,39–41).
Corticosteroid Properties
Corticosteroids are powerful anti-in-
flammatory medications. The rationale
for administration is the suppression
of inflammation implicated in the path-
ogenesis of radiculopathy and axial
pain (42). Inflammation as a generic
physiologic response can be triggered
by numerous stimuli. In disk hernia-
tion and spondylosis, radiculopathy re-
sults from both chemical and mechan-
ical irritants (43–45). Phospholipase
A2 and other enzymes are released
into the epidural space by disk mate-
rial and annular tear. These inflamma-
tory mediators recruit macrophages
that secrete cytokines and catalyze the
inflammatory cascade; they produce
prostaglandins and leukotrienes that
sustain the inflammatory cycle (46).
Mechanical nerve root stretching,
tethering, and compression provoke
Published online
10.1148/radiol.2016152055  Content codes:
Radiology 2016; 281:669–688
Abbreviations:
ESI = epidural steroid injection
FDA = Food and Drug Administration
NRB = nerve root block
Conflicts of interest are listed at the end of this article.
HOW I DO IT: Spinal Injections for Pain Management Palmer

the same inflammatory reaction due Figure 1

to cellular damage and ischemic in-
jury (47). Corticosteroids inhibit the
release of cytokines and the synthesis
of prostaglandins, thereby suppressing
the physiologic response (42,48).
Injectable corticosteroids are non-
particulate or particulate. Nonpartic-
ulate corticosteroids have rapid-onset
but short-lived anti-inflammatory
effects, whereas particulate cortico-
steroids have delayed but sustained
effects. Nonparticulate preparations,
including dexamethasone sodium
phosphate and betamethasone
sodium phosphate, are fully soluble
and clear in appearance. Particulate
suspensions contain corticosteroid
esters that are insoluble in saline, lo-
cal anesthetic, and iodinated contrast
material (49). Particulate corticoste-
roids, including methylprednisolone
acetate, triamcinolone acetonide,
and betamethasone acetate, create a
white cloudy mixture when agitated. Figure 1:  Schematic drawing of the prone lumbar spine shows lumbar structures, needle placements, and
Methylprednisolone has the largest cross-sectional relationships of structures in the spinal canal and foramina. Radicular arteries may be paired,
particles, whereas betamethasone has unilateral, or absent. In lumbar NRB, the needle targets the epiradicular space that surrounds the ventral ra-
the smallest (49–51). mus of the spinal nerve, the dorsal root ganglion (∗), and the nerve roots. In ESI, the needle targets the small
Particulate corticosteroids have triangle of fat in the dorsal epidural space. In lumbar facet injection (F1), needle trajectory must account for
been associated with neurologic dam- joint curvature. Otherwise, joint access is blocked by margins of articulating processes (F2).
age during transforaminal cervical and
lumbar injections (51–54). Because of been a widespread practice for many
Interaction with Patients
limited data, the incidence of these decades; however, the effectiveness
rare complications cannot be deter- and safety of the drugs for this use Structured interactions promote trust
mined (55). The proposed mechanism have not been established, and FDA and build patient-physician relation-
is direct intra-arterial administration has not approved corticosteroids for ships that can last for years in individ-
resulting in embolic occlusion followed such use.” Several professional soci- uals who suffer from chronic neck or
immediately by ischemia or infarction eties responded to this FDA safety back pain and who return for periodic
of neural tissue (Figs 1, 2). Neurologic announcement, expressing concern injections. In my practice, these inter-
sequelae, including quadriplegia, para- that the FDA warning did “not differ- actions are important because patients
plegia, and death, are extremely rare entiate between the risks and benefits relay their positive and negative expe-
but catastrophic complications, caus- of transforaminal versus interlaminar riences to referring physicians. These
ing some authors to advocate use of routes of administration, and partic- interactions also create rewarding op-
a nonparticulate corticosteroid in all ulate versus non-particulate formula- portunities for radiologists to counsel
transforaminal procedures (50,56,57). tions of steroids” (60). patients and affect clinical decision
No neurologic complications due to Numerous factors influence the making. Patients actively seek the per-
nonparticulate corticosteroid use have choice of injected corticosteroid, in- spective of nonsurgeons. Radiologists
been reported (58). cluding FDA warnings, published arti- should understand the treatment op-
In April 2014, the Food and Drug cles, recommendations from specialty tions and lifestyle modifications that
Administration (FDA) posted a safety societies, and package inserts. Data help patients stay active.
announcement requiring manufac- from these varied sources may be During procedural visits, I engage
turers to add a warning in package in- contradictory, requiring judgments on patients at four junctures to obtain or
serts about adverse neurologic events safety and efficacy based on personal convey information. These purposeful
(59). In that same posting, the FDA experience, the regulatory environ- interactions can be categorized as the
stated: “Injecting corticosteroids into ment, and the availability of drugs in interview, the blow-by-blow, the teach-
the epidural space of the spine has the hospital formulary. able moment, and the discharge. The

Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 671

HOW I DO IT: Spinal Injections for Pain Management Palmer

Figure 2
Figure 2:  Cervical structures and needle placement
for NRB. (a) Schematic drawing of the supine cervical
spine depicts cross-sectional relationships of structures
in the spinal canal and foramina. Radicular and segmen-
tal medullary arteries are too small to identify with MR
imaging. In cervical NRB, needle placement is pos-
terior to great vessels and nerve plexus. The needle
targets the foramen posteriorly and inferiorly in close
proximity to the exiting nerve and dorsal root ganglion
(∗). AS = anterior scalene muscle, CA = carotid artery,
JV = jugular vein, MS = middle scalene muscle. (b)
CT angiographic image at C5–6 disk level shows
posterior location of the jugular vein (JV). A graphically
depicted needle (NRB) is overlaid at 45° and shows
superimposition on the external jugular vein (EJ). To shift
the great vessels from the needle path during cervical
NRB, turn the head away from the symptomatic side.
Lateral approach may be necessary when MR imaging
reveals great vessels in a posterior location. Enhancing
vessels (arrows) surround nerve root ganglia (∗) in
foramina. CA = carotid artery, VA = vertebral artery.

first interaction, the interview, focuses
on clinical history and incorporates the
consent process. The second and third
interactions, the blow-by-blow and the
teachable moment, respectively, take
place during and immediately after the
intervention while the patient lies on
the fluoroscopy table. The final interac-
tion, the discharge, occurs once the pa-
tient is dressed and is waiting to leave.
The Interview
The patient interview is critical in
procedural selection and planning.
I have four goals: (a) obtain a fo-
cused clinical history, (b) correlate
symptoms with imaging findings, (c)
approve the requested procedure or
propose a different one, and (d) ob-
tain written informed consent. In new
patients, these goals often can be ac-
complished within 10 minutes. In re-
turning patients, less time is required
because usually little has changed and
previously successful procedures can
be repeated. It only takes a few ques-
tions to determine the outcome of the
prior injection and reestablish the
pain generator.
During history taking, one should
focus on spine-related symptoms. Sys-
tematic questioning (Fig 3) quickly
yields enough clinical information to
guide the targeted inspection of imag-
ing studies and the formulation of a
treatment plan (correlation of symp-
toms and imaging findings is addressed
in the next section). Begin with the
discrimination of axial pain from ra-
diculopathy. When leg and back pain
coexist, ask which one is predominant.
Most patients would be willing to live

672 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016

HOW I DO IT: Spinal Injections for Pain Management Palmer
Figure 3
Figure 3:  Flowchart shows interview questions to be asked during history
taking. A = answer, Q = question.
with the lesser pain if the major pain
could be relieved.
Radicular symptoms often enable
one to verify the pain generator during
imaging correlation. In contrast, non-
specific axial pain poses diagnostic chal-
lenges. It can be acute or chronic, mild
or severe, intermittent or constant, dull
or sharp, localized or migratory, or any
Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 673
combination thereof. Patients often
point to a general region in the neck
or low back. Clinical history and phys-
ical examination have limited value in
determining the cause of axial pain and
guiding procedural selection (61).
Informed consent follows his-
tory taking, correlation of symptoms
with imaging findings (hereafter,
symptom-imaging correlation), and
procedural selection. Explain the pro-
cedure, and educate the patient about
risks. It may be necessary to disclose
that the FDA has not approved cortico-
steroids for epidural administration. Be-
sides fulfilling ethical and regulatory re-
quirements, the consent process should
uncover clinical conditions, medications,
and allergies that increase risk and com-
promise outcome. The major issues are
anticoagulation therapy, active infection,
and contrast material reaction. When
you are describing bleeding and infec-
tion risks, ask about anticoagulant and
antibiotic treatments. Inquire about la-
tex allergy, diabetes, and other recent
steroid injections. In my practice, our
administrative assistant screens patients
prior to the procedure to avoid the dis-
covery of unresolvable issues, such as
anticoagulation therapy, in the fluoros-
copy suite.
The Blow-by-Blow
Occasionally, patients refuse to receive
information during the procedure. As a
coping strategy, they wear headphones
to listen to music, or they prefer silence,
choosing to mentally transport them-
selves to another place. However, most
patients want to know what is happening
and value a blow-by-blow narrative. The
blow-by-blow serves several purposes.
Patients feel connected and informed.
It relieves anxiety and eliminates the
element of surprise. It conveys forward
progress. Verbal communication also en-
gages the technologist, fellow, and any
other individuals involved in the proce-
dure. My custom is to announce when I
am deciding where to insert the needle,
putting a dot on the skin with a marker,
cleaning the skin, preparing the medica-
tions, numbing the skin, positioning the
needle, and injecting dye to make sure
the needle is in the right place. I express
my satisfaction with needle placement
before I inject the steroid solution. Dur-
ing injection, I warn patients that the
injection could cause pressure or pain.
When concordant symptoms are pro-
duced, I reassure patients and state that
the needle is correctly placed. Patients
appreciate knowing that the procedure
is going as expected.
HOW I DO IT: Spinal Injections for Pain Management Palmer
The Teachable Moment
The teachable moment begins immedi-
ately after needle removal while wash-
ing off the skin antiseptic. To set pos-
itive expectations, state that the goal
of the procedure was achieved (ie, the
steroid was delivered to the intended
target). Indicate the rationale for inject-
ing the steroid. Explain that the steroid
works by decreasing inflammation, not
by shrinking the disk herniation, re-
versing arthritis, or opening stenotic
spinal canals. Patients should under-
stand that the drug is a powerful an-
ti-inflammatory agent but that the de-
gree of pain relief depends on whether
inflammation is causing the symptoms.
In patients who might benefit from see-
ing the fluoroscopic images, reinforce
the technical success of the procedure
by pointing out needle placement and
contrast material flow on the monitor.
The Discharge
The discharge process generates in-
formation about immediate pain re-
sponse. Symptoms might be decreased,
unchanged, or increased depending on
the level of preprocedural pain and the
volume of injected anesthetic. Prompt
pain relief creates a positive attitude
about the procedure and promotes the
placebo effect. A surprising number
of patients claim pain reduction even
if no local anesthetic was injected. In
dictated reports, record the postproce-
dural pain response (eg, right leg pain
decreased from a score of 8 of 10 to
a score of 2 of 10). If symptoms are
already improved at the time of dis-
charge, I continue to set positive expec-
tations by explaining to the patient that
the steroid was mixed with anesthetic
and, therefore, it is in the same correct
location.
One must explain the time frame
for steroid effectiveness and provide
activity guidelines. Patients can become
disappointed the day after injection if
their pain remains unchanged. Because
particles release the steroid gradually,
it may take 12–24 hours for the drug
to take effect, 4–6 days for its effects
to become more pronounced, and
more than a week for it to reach full
effectiveness. Advise patients to limit
674 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016
themselves to baseline levels of exercise
and physical therapy for 4–6 days. Pa-
tients whose condition improves after
2–3 days are tempted to overdo it be-
fore the drug has reached full effective-
ness, thereby stirring up inflammation
that overwhelms the steroid and dimin-
ishes the overall treatment benefit.
Patients often ask how long the
injection will help. The time course is
surprisingly predictable in patients with
chronic conditions, such as spinal ste-
nosis and facet arthropathy. Symptoms
decrease during the first 2–3 weeks
after injection when the anti-inflamma-
tory effects are strongest but return to
baseline levels over the following 6–8
weeks as the particulate steroid dissi-
pates. In patients with acute conditions,
such as disk herniation and annular
tear, the steroid can break the inflam-
matory cycle and relieve pain for more
than 6–8 weeks. When new symptoms
are superimposed on long-standing
ones, such as acute radiculopathy su-
perimposed on chronic low back pain,
explain that corticosteroid injection
may accelerate a return to the baseline
condition. Steroid administration de-
creases the new reversible nocioceptive
pain but leaves the long-standing irre-
versible neuropathic pain unchanged.
Role of Imaging in Procedural Selection
Symptom-imaging correlation guides
procedural selection and planning
(Movie 1 [online]). It enables one to
verify the appropriateness of the re-
quested intervention or justify modifi-
cation. Procedural modification is most
practical when the radiologist has au-
thorization to proceed independently.
For the radiologist who possesses the
skill, experience, and confidence to
assume responsibility for treatment
decisions and, therefore, therapeutic
outcomes, the role in pain management
expands beyond rote injection.
Imaging results are available in the
majority of cases. During screening, our
administrative assistant asks patients to
bring their MR images if they were ob-
tained at a different institution. On rare
occasions, we proceed without the bene-
fit of symptom-imaging correlation.
Procedural selection begins with a
judgment on the pain generator. The
following symptom-imaging correlation
exercise is popular with our fellows
because it hones skills in both history
taking and MR image interpretation.
The exercise has two variations. In
one variation, we interview the patient
before we review the MR images. We
then deduce the most likely pain gener-
ator and predict the MR imaging find-
ings. In the other variation, we review
the MR images before we interview
the patient, then we deduce the most
likely pain generator and predict the
patient’s symptoms. Symptom-imaging
correlations are often obvious, but sur-
prising mismatches do occur. These
mismatches teach valuable nuances
in pain management and MR image
interpretation.
In younger patients with acute or
subacute radiculopathy, dermatomal in-
formation serves to focus MR image re-
view. Symptoms usually correlate per-
fectly with nerve entrapment because
of lateralization of single-level disk ab-
normalities. Occasionally, a symptom-
specific search will lead to the diagno-
sis of an intraforaminal or lateral disk
extrusion that explains symptoms but
that was overlooked at the time of MR
image interpretation (Fig 4). Transfo-
raminal NRB targets the pain generator
and delivers the steroid directly to the
inflamed nerve root (62–64) (Figs 5, 6).
In older patients with chronic uni-
lateral radiculopathy, symptom-imaging
correlation is more challenging because
of multilevel spondylosis. Therapeu-
tic success is also more challenging
when severe stenosis causes irrevers-
ible nerve damage and neuropathic
pain. Transforaminal NRB remains a
recommended treatment option if der-
matomal information reveals a specific
pain generator. One exception is nerve
compression by a facet cyst. To address
this problem, one can combine percuta-
neous cyst rupture with intra-articular
corticosteroid injection (65) (Fig 7).
In older patients with chronic bi-
lateral radiculopathy, the radiologist
should solicit signs of neurogenic clau-
dication. Intermittent back pain is pre-
cipitated by prolonged standing or
HOW I DO IT: Spinal Injections for Pain Management Palmer

Figure 4 Figure 5

Figure 5:  Lumbar NRB in a 64-year-old

man with right leg pain correlating with
Figure 4:  Intraforaminal disk extrusion missed at prospective interpretation in
L4–5 lateral recess stenosis and right L5 me-
a 41-year-old woman with acute L4 radiculopathy. Axial T2-weighted MR image
chanical impingement. Anteroposterior fluoro-
shows unsuspected L4–5 intraforaminal disk extrusion (white arrowhead)
scopic image in the prone position shows the
causing mechanical impingement on left L4 nerve root ganglion (arrow). During
needle (arrow) in supraneural location between
NRB (not shown), the needle targeted the L4 ventral ramus peripheral to the
L5 and S1 pedicles (∗). Contrast material
foramen to avoid severe pain production due to foraminal stenosis and nerve
(arrowheads) flows cranially into the spinal canal
root displacement. Right L4 nerve root ganglion (black arrowhead) is in its
along the L5 epiradicular space to the level of
normal location and is surrounded by fat.
the L4–5 disk space and pain generator in lateral
recess.
walking and is relieved by sitting or predict treatment responses to facet in-
leaning forward. One should expect MR jections (73). If corticosteroid adminis-
imaging to reveal spinal stenosis, which tration alleviates symptoms, systematic procedural alternatives are ESI and fac-
is the leading reason for spinal surgery anesthetic injections (medial branch et injection. These limited choices seem
(18). When MR imaging shows multi- blocks) yield corroborative diagnostic trivial enough, but decision making can
level stenosis, dermatomal information information prior to radiofrequency be problematic due to overlapping clin-
may indicate the level of pain generator. ablation. ical syndromes and equivocal MR im-
Interlaminar ESI is the recommended Segmental instability creates multi- aging findings (66–68). Incidental MR
procedure because the corticosteroid ple pain generators and causes debil- imaging abnormalities in asymptomatic
can spread cranially and caudally over itating symptoms that respond poorly adults are difficult to discriminate from
multiple disk levels. to injections. Initially, when back pain true pain sources in symptomatic pa-
When symptoms suggest lum- predominates, symptoms may respond tients (15,61,69,70). The default inter-
bar facet syndrome (posterior ramus to ESI, facet injections, or a combina- vention is ESI. In the absence of MR im-
syndrome), one must scrutinize the tion thereof. Progressive facet degener- aging findings that support a different
zygapophyseal joints for signs of inflam- ation leads to articular hypermobility, level of injection, the radiologist should
mation, including effusion, capsulitis, attritional bone loss, and malalign- select the L4–5 level because injectate
and periarticular edema. Back pain ment. Increasing anterolisthesis exac- typically flows cranially and caudally,
may radiate into the buttocks, groin, erbates spinal stenosis and foraminal and it will cover the three lowest mov-
or posterior thigh and may worsen with nerve impingement. When neurogenic able disk spaces.
prolonged standing and extension and claudication and radiculopathy become One should inspect MR images for
rotation or lateral bending movements superimposed on back pain, segmental bone marrow edema involving facets,
(71). Sclerotomal maps for posterior instability often forces surgical inter- pars defects, spinous processes (Baas-
rami depict the patterns of referred vention and spinal fusion. trup syndrome), and lumbosacral pseu-
pain from facet joints but are less ac- In patients with chronic nonlocal- doarthroses (Bertolotti syndrome).
curate than dermatomal maps for pat- izing low back pain, one must inspect When symptoms are highly localized,
terns of referred pain from ventral rami disks and facets to judge the rela- ask the patient to point to the most
(72). Clinical history and physical ex- tive importance of discogenic and ar- painful spot and document that site
amination findings cannot be used to thropathic abnormalities. The basic fluoroscopically, including the pointing

Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 675

HOW I DO IT: Spinal Injections for Pain Management Palmer

Figure 6

Figure 6:  Cervical NRB in 41-year-old woman with left arm pain correlating with C6–7 disk extrusion and left C7 mechanical impingement. (a) Oblique fluoroscopic
image obtained in the supine position shows foramina (curved lines) are opened by lateral rotation of the detector approximately 45°. During set-up, turn the head
away from the side of needle placement, thereby displacing the great vessels and nerve plexus from the needle path. The needle tip (arrow) is directed inferiorly
toward C6–7 foramen to target exiting left C7 nerve and posteriorly to avoid vertebral artery. The needle hub has been preloaded with contrast material (arrowhead)
to obviate gas injection. (b) Digital subtraction posteroanterior fluoroscopic image shows contrast material (arrowheads) flowing along the epiradicular space into the
foramen between C6 and C7 pedicles (∗). Subtraction technique may improve detection of intravascular contrast material. The needle (arrow) is barely visible. (c)
Subsequent posteroanterior fluoroscopic image shows contrast material (arrowheads) spreading along C7 nerve between pedicles (∗) into the epidural space. The
needle (arrow) terminates at the border of lateral masses. It can be advanced into outer foraminal thirds if contrast material flow is unsatisfactory (intravascular or
extraforaminal).

Figure 7

Figure 7:  Facet cyst rupture in 71-year-old man with left L5 radiculopathy correlating with left L5-S1 foraminal cyst. (a) Axial T2-weighted MR image at L5-S1
level shows foraminal cyst (straight arrow) is contiguous with degenerated left L5-S1 facet joint (curved arrow) and proximate to left L5 nerve (arrowhead). (b)
Anteroposterior fluoroscopic image in the prone position shows contrast material in the needle hub (straight arrow) and facet cyst (curved arrow). The inferior recess
of the left L5-S1 facet joint was accessed with anteroposterior fluoroscopy by moving the needle caudally off the articulating process. (c) Subsequent anteroposterior
fluoroscopic image shows epidural flow of contrast material (arrowheads), indicating cyst rupture. Contrast material no longer fills the cyst or needle hub (arrow).
Corticosteroid may prevent or delay recurrence of facet cyst.

finger. If this site corresponds to bone
marrow edema or another potential
pain generator on MR images, it can be
targeted for diagnostic information and
therapeutic response.
Pearls and Pitfalls of Fluoroscopy-
guided Injections
The following pearls and pitfalls focus
on fluoroscopy-guided injections (Movie
2 [online]). Expert interventionalists
develop individualized techniques and
often approach the same problems and
procedures in different ways. Some
physicians prefer to use computed

676 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016

HOW I DO IT: Spinal Injections for Pain Management Palmer
tomography (CT) to guide needle place-
ment (74–80). This preference reflects
training experience, resource availabil-
ity, and institutional policy. No studies
have compared the effectiveness or
safety of CT-guided interventions versus
fluoroscopy-guided interventions. Neu-
rologic complications have occurred
during spine injections performed with
both modalities (81–84).
Advantages of CT include higher
reimbursement and cross-sectional
documentation of needle position. CT
may improve the accuracy and safety
of needle placement when overlapping
bones make it difficult to plan an unob-
structed needle trajectory during fluo-
roscopic set-up and when intervening
soft-tissue structures must be avoided
but cannot be seen fluoroscopically. In
thoracic NRB, for example, CT shows
the complex relationship between the
ribs, transverse processes, and lung.
For atlantoaxial (C1–2) facet injection,
CT helps to avert complications from
arterial puncture by depicting the ver-
tebral artery (82). In lumbar facet in-
jection, when intraarticular placement
is essential for diagnostic information
or therapeutic cyst rupture, CT shows
marginal osteophytes and hairline joint
spaces (85).
Disadvantages of CT include higher
cost, increased radiation dose, longer
procedural time, and lack of availabil-
ity. The radiation dose of CT interven-
tions can be decreased to the level of
fluoroscopic interventions by using a
low-dose protocol that eliminates ac-
quisition of a topogram, minimizes both
energy and tube current, and severely
restricts the number of image acquisi-
tions during needle placement (86–88).
At institutions where interventional
CT is a limited resource, CT-guided in-
jections for pain management can be
challenging to schedule if they must
compete for scanner time with biopsy,
oncologic ablation, and abscess drain-
age procedures.
An advantage of fluoroscopy is the
live real-time observation of contrast
material flow and, therefore, vessel
opacification in the case of inadver-
tent intravascular needle placement.
During CT-guided procedures, because
Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 677
conventional scanning is delayed until
operators leave the room, intravascular
contrast material washes away, thereby
precluding vessel identification. Inter-
mittent fluoroscopy, if performed before
and after but not during contrast mate-
rial injection, also fails to show opaci-
fied vessels (52,89). In mixed injections
with concurrent intra- and extravascu-
lar contrast material flow, only the ex-
travascular contrast material remains
visible, creating the false reassurance of
extravascular needle placement. Some
authors have proposed the use of CT-
fluoroscopic techniques to improve the
real-time detection of opacified vessels
(90–92). However, vessels outside of
the limited stack of CT images remain
impossible to identify.
A potential advantage of fluoros-
copy is the range of detector rotation,
which enables steep craniocaudal angu-
lation. Steeper craniocaudal angles are
often required in L5 and S1 NRBs, as
well as in lumbar ESIs, in the setting
of interlaminar collapse or exaggerated
lumbar lordosis.
The advantages of both modal-
ities can be attained with one unit
that combines C-arm fluoroscopy with
cone-beam CT (93,94). The flat-panel
detector spins and acquires a volu-
metric data set enabling multiplanar
two-dimensional reformations and
three-dimensional reconstructions. Be-
cause the fluoroscopic image is over-
laid onto the three-dimensional data
set, the unit can align itself accord-
ing to selected skin entry and target
locations for bull’s-eye needle naviga-
tion. If necessary, final needle posi-
tion and contrast material location are
documented with a second volumetric
acquisition.
Risk and Risk Mitigation
Adverse events are exceedingly rare
when experienced practitioners use
fluoroscopic guidance and inject con-
trast material to confirm needle posi-
tion (16). In more than 8000 cervical,
thoracic, and lumbar interventions per-
formed by me or under my supervision,
none have been complicated by hem-
orrhage, infection, or neurologic dam-
age. Radiologists should recognize and
manage immediate and delayed com-
plications or perform patient triage for
appropriate care. Adverse events can
occur during injection (pain, hemor-
rhage, reaction to contrast material, va-
sovagal reaction, dural puncture, nerve
or vessel damage), immediately after
injection (pain, hemorrhage, extremity
weakness, paresthesia), or days later
(infection, headache, flushing reaction
to steroid). Most adverse events can be
avoided by anticipating risks discovered
during history taking and image review.
Bleeding risk increases with age,
underlying coagulopathy, severity of
spondylopathy, and difficulty of needle
placement (95). Although the incidence
is unknown, bleeding risk increases in
patients who have undergone anticoagu-
lation therapy, and it increases substan-
tially in patients taking multiple antico-
agulant and antiplatelet medications,
including nonsteroidal anti-inflamma-
tory drugs (95). Epidural hematoma
rarely occurs; however, it poses the
greatest threat because of spinal cord
or cauda equina compression, and it
requires surgical evacuation to prevent
permanent neurologic sequelae. Inci-
dence has been estimated at 1:220 000
after subarachnoid anesthesia and at
1:150 000 after epidural anesthesia in
healthy patients (96). Epidural hema-
toma has been described after ESI and
facet injection in patients without co-
agulopathy or anticoagulation therapy
(96). Patients should discontinue use
of anticoagulants for appropriate inter-
vals, and they should coordinate bridg-
ing therapy according to instructions
from referring physicians or consulting
cardiologists (97). ESI is considered
safe in patients taking nonsteroidal an-
ti-inflammatory drugs (98). In a study
of 1214 patients who underwent ESI,
no hemorrhagic complications occurred
in 383 (32%) patients taking nonsteroi-
dal anti-inflammatory drugs (98).
Iodinated contrast material should
be approved for myelography in case of
inadvertent intrathecal administration.
One should recognize patterns of layer-
ing subarachnoid contrast to avoid sad-
dle anesthesia and ascending paralysis
from anesthetics and arachnoiditis from
corticosteroids. Nearly 3% of scheduled
HOW I DO IT: Spinal Injections for Pain Management Palmer
Figure 8
Figure 8:  Intravenous injection during interlaminar
ESI in a 48-year-old woman with axial low back
pain after L4–5 anterior fusion. Lateral fluoroscopic
image obtained during contrast material injection
shows needle tip (arrow) projecting over the spinal
canal at L3–4 disk level. Extensive epidural venous
network (arrowheads) is visible from L2 to L5.
Needle repositioning could not evade vessels. Thus,
ESI was performed at L2–3 (images not shown).
∗ = L4–5 interbody cage.
patients have had known or suspected
reactions to contrast material (16). In
these patients, options include (a) pre-
medication with oral prednisone and
diphenhydramine, (b) steroid injection
without contrast material confirmation
of needle location, and (c) injection
of a gadolinium-based contrast agent.
Gadolinium chelates provide off-label
alternatives to iodinated contrast mate-
rial and appear safe for use in epidural
injection (99). Intrathecal administra-
tion should be carefully avoided. Digital
subtraction fluoroscopy may improve
visualization of gadolinium-based con-
trast material, which is less radiopaque
than iodinated contrast material.
Extensive arterial and venous net-
works crisscross the epidural and epi-
radicular spaces of the spinal canal and
neural foramina (Figs 1, 2, 8). Overall,
678 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016
intravenous needle placements are
more common in patients with cervical
NRBs (incidence range, 19.4%–32.8%)
than in those with lumbar NRBs (inci-
dence range, 11.2%–13.1%), but actual
incidence depends on injection level
(10–13). In the cervical spine, vascular
cannulation is more likely from C3–4
to C5–6 (range, 40%–57%) (13). In
the lumbosacral spine, it is more than
twice as likely at S1 (21.3% incidence)
compared to lumbar levels (8.1% in-
cidence) (10–12). In both the cervical
spine and the lumbar spine, aspiration
fails to produce a flashback of blood in
45%–73% of intravenous needle place-
ments subsequently proved via contrast
material opacification (10,100).
Simultaneous intra- and extravascu-
lar contrast material flow (mixed injec-
tion) is observed more commonly than
is intravascular flow alone in both cer-
vical NRBs and lumbar NRBs (12,13).
In the cervical spine, mixed injection
was reported in 18.9% of injections
as compared with vascular flow alone,
which was reported in 13.9% of in-
jections (13). Concurrent flow creates
challenges because extravascular con-
trast material can obscure vessels. To
best detect vessels, one should optimize
fluoroscopic techniques by decreasing
the field of view, dimming the lights, us-
ing digital subtraction angiography, and
injecting contrast agents that contain a
higher concentration of organic iodine
(eg, 300 mg/mL).
Injected steroids have systemic glu-
cocorticoid effects in addition to local
anti-inflammatory effects. Fortuitous
benefits include temporarily decreased
pain from arthritis and spondyloar-
thropathy. Undesired consequences
include elevation of the blood glu-
cose level, suppression of the immune
system, and suppression of the hypo-
thalamic-pituitary axis. One should
caution patients with diabetes to mon-
itor their blood glucose levels for 7–10
days after the procedure. Ask patients
about antibiotics they are taking, and
reschedule patients who are taking an-
tibiotics for active infections. Prophy-
lactic antibiotics are not a contraindi-
cation. Repeated steroid injections at
short intervals (less than 8 weeks for
particulate preparations) can inhibit
recovery of the hypothalamic-pituitary
axis and can lead to decreased bone
mineral density (101–103). Cushingoid
symptoms, although rare, can persist
for several months after steroid injec-
tions have been terminated (28,104).
Injectate composition and volume
depend on numerous factors and vary
widely between practitioners (Table).
Choice of corticosteroid should take
into account the risk of intravascular
or intrathecal injection and, there-
fore, particle size and the addition of
preservatives (benzyl alchohol) or ve-
hicles (polyethylene glycol) (105,106).
Corticosteroid particles may aggregate
and form larger particles when mixed
with local anesthetics and contrast
agents containing certain preservatives
(49,50).
Commonly used anesthetics include
lidocaine, bupivacaine, and ropiva-
caine. Formulations with preservatives
(methylparaben) and vasoconstrictors
(epinephrine) should be avoided. Meth-
ylparaben is classified as an antimicro-
bial preservative, and it is added for its
bacteriostatic activity. Choice of anes-
thetic should take into account patient
health, pain severity, and postdischarge
activities. Use anesthetics with caution
in elderly or unsteady patients and in
individuals who are planning to drive or
take public transportation immediately
after they are discharged. Patients with
baseline weakness or paresthesia are
more susceptible to anesthetic effects
and may develop profound postproce-
dural weakness, even if only a small vol-
ume of anesthetic is administered.
Procedural Tips and Techniques
In the majority of spine procedures, pa-
tients are placed in the prone position.
This position can exaggerate lumbar
lordosis, aggravate nerve root entrap-
ment, and provoke facet-related pain.
Patient comfort is more important than
perfect prone positioning, as the goal
is to limit progressive discomfort and
involuntary movement. Placement of
a pillow under the pelvis can help to
reduce lordosis and can relieve symp-
toms. Patients who cannot lie in the
prone position may be able to tolerate
HOW I DO IT: Spinal Injections for Pain Management Palmer

Drug Doses in Common Therapeutic Spinal Injections

Injection Type Dose

Lumbar ESI
  MGH protocol* BTM 15 mg (2.5 mL) followed by saline (2–4 mL) mixed with lidocaine 1% 0–10 mg (0–1 mL)
  Published protocol†
  Corticosteroid dose MPA 40–120 mg (1–3 mL), TCA 60–120 mg (1.5–3 mL), BTM 9–18 mg (1.5–3 mL)
   Anesthetic dose Bupivacaine 0.25%–0.5% 15–40 mg (3–8 mL), lidocaine 0.5%–1.0% 15–40 mg (3–8 mL)
  Saline volume Saline 0–8 mL
Thoracolumbar transforaminal nerve root injection
  MGH protocol at or above conus DSP 8–12 mg (2–3 mL)
  MGH protocol below conus‡ BTM 6–15 mg (1–2.5 mL) mixed with saline 1–2.5 mL & lidocaine 1% 0–10 mg (0–1 mL)
  Published protocols
  Corticosteroid dose MPA 20–80 mg (0.5–2.0 mL), TCA 20–40 mg (0.5–1.0 mL), BTM 6–9 mg (1.0–1.5 mL), DSP 4–8 mg (1–2 mL)
   Anesthetic dose Lidocaine 0.5%–1.0% 10–20 mg (1–4 mL), bupivacaine 0.25%–0.5% 5–20 mg (1–8 mL)
Cervical transforaminal nerve root injection
  MGH protocol DSP 8–12 mg (2–3 mL)
  Published protocols
  Corticosteroid dose DSP 4–12 mg (1–3 mL), BTM 6 mg (1 mL), TCA 40 mg (1 mL), MPA 40 mg (1 mL)
   Anesthetic dose Bupivacaine 0.25%–0.5% 2.5–5 mg (1 mL), lidocaine 0.5%–2.0% 5–20 mg (1 mL)
Lumbar or cervical facet injection
  MGH protocol TCA 40 mg (1 mL) mixed with ropivacaine 0.5% 5 mg (1 mL) or lidocaine 1% 10 mg (1 mL)
  Published protocols
  Corticosteroid dose TCA 20–60 mg (0.5–1.5 mL), BTM 3–6 mg (0.5–1.0 mL), MPA 20–60 mg (0.5–1.5 mL)
   Anesthetic dose Lidocaine 1%–2% 5–20 mg (0.5–1.0 mL)

Note.—Medication and dose selections require physician discretion. BTM = betamethasone acetate, 3 mg/mL and betamethasone sodium phosphate, 3 mg/mL; DSP = dexamethasone sodium
phosphate, 4 mg/mL; MPA = methylprednisolone acetate, 40 mg/mL; TCA = triamcinalone acetonide, 40 mg/mL.
* Massachusetts General Hospital (MGH) protocols are routinely modified based on recommendations by specialty societies and consensus groups.
†
Published protocols may not reflect current recommendations by specialty societies and consensus groups.
‡
Corticosteroid dose is proportionate to epidural flow.

imaging in the oblique or lateral decu-
bitus position.
Fluoroscopic set-up can neutralize
the challenges posed by lordosis, sco-
liosis, spondylolisthesis, and oblique
patient positioning. By standardizing
needle trajectory, even deformed spon-
dylotic spines can be depicted in con-
ventional anteroposterior, lateral, and
oblique views at the level of injection.
Set-up usually involves four systematic
maneuvers. First, one must identify the
level of intervention. Second, one must
rotate the detector left-right to obtain
a straight anteroposterior projection of
the spine. Third, one must adjust the
craniocaudal tilt to normalize endplate
relationships. The final maneuver in-
volves rotating the detector left-right to
obtain a bull’s-eye projection for needle
navigation. Fluoroscopy units that en-
able one to use saved positions allow
rapid transition between different views.
One should archive fluoroscopic
images to document the procedure,
needle position, and contrast material
distribution for billing and medicolegal
purposes. Referring surgeons may re-
view the images with patients and de-
termine whether the steroid reached its
intended target. In patients who have
undergone repeated injections, the im-
ages provide templates for reproducing
safe and effective needle placement.
One should archive a minimum of two
images. The first image, obtained be-
fore steroid delivery, shows extravas-
cular contrast material. The second,
obtained after steroid delivery, shows
contrast material washout and proves
that the steroid flowed into the same
location as the contrast material.
Lumbosacral transforaminal injec-
tion.—The rationale for transforami-
nal NRB is precise drug delivery to the
inflamed nerve root (Figs 5, 9). The
epiradicular space, which is the target
in needle placement, surrounds the
ventral ramus. Because the epiradicu-
lar and epidural spaces are continuous,
injectate flows selectively along the spi-
nal nerve and nerve root into the spinal
canal (Figs 5, 6). Transforaminal injec-
tion delivers the corticosteroid directly
to the pain generator in the foramen
or ventral epidural space. Thus, NRB
can yield a superior therapeutic effect
with a smaller corticosteroid dose than
that used with interlaminar ESI (35,62–
64,107). In ESI, the corticosteroid takes
the path of least resistance, spreading
indiscriminately from the dorsal epidu-
ral space throughout the spinal canal.
When the foramen is patent, intra-
foraminal needle placement can be su-
praneural (subpediculate) or infraneural
(retrodiskal) (Figs 10, 11). Select the
supraneural location, the so-called safe
triangle, for reliable epiradicular flow of

Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 679

HOW I DO IT: Spinal Injections for Pain Management Palmer

Figure 9 foramen. Do not rotate a curved needle

in the foramen because of the risk of
lacerating vessels, including the lateral
sacral artery. In patients with osteope-
nia, a 22-gauge straight needle can be
used to penetrate sacral plates with a
twisting or oscillating motion. Appro-
priate epidural depth is determined
with lateral fluoroscopy. Trajectory can-
not be altered once the needle is drilled
through bone. The same transosseous
technique can be used to advance a
straight needle through a paraspinal fu-
sion mass for lumbar NRB.
During most NRBs, expect injection
to produce transient radicular symp-
toms. To avoid the overproduction of
severe long-lasting pain, ask patients
to control the injection rate. Explain
that symptoms should not exceed 5–6
Figure 9:  Fluoroscopic set-up for sacral (S1) NRB in a 53-year-old man with left leg pain correlating with on a 0–10 pain scale. Patients indicate
L5-S1 disk extrusion and left S1 mechanical impingement. (a) Anteroposterior fluoroscopic image shows the when to inject and when to pause. Pa-
detector was tilted craniocaudally to align the inferior margin of posterior S1 foramen (thin curved line) with tients are reassured by this degree of
the superior margin of anterior S1 neural arch (arrowheads). Sacral orientation determines the degree of control; however, stoic individuals may
craniocaudal tilt. The detector was rotated laterally to align the medial margin (thick curved line) of S1 pedicle request continued injection despite
(∗) with lateral margin of posterior S1 neural foramen along expected course of S1 nerve root. Curved needle severe pain because they fear partial
(arrow) improves foraminal navigation. Until it enters the foramen, the needle must target the inferolateral steroid dosing. Watch their faces for
border of the posterior S1 foramen. (b) Subsequent anteroposterior fluoroscopic image shows the needle signs of discomfort, and observe their
hub (white arrow) and needle tip (white arrowhead) are oriented cranially along the expected course of the body language. If the injection rate is
S1 nerve root. Initially, contrast material flowed retrograde into the posterior S1 foramen (black arrow). After too slow, steroid particles can settle
advancing the needle, contrast material (black arrowheads) spread favorably along S1 pedicle (∗) and S1 in the needle and clog it. When injec-
nerve root to the L5-S1 disk level. tion pressure unexpectedly increases,
reinsert the stylet to clear the needle
the steroid along the nerve root to the Infraneural placement is more likely to before a steroid plug completely blocks
next higher disk level (64) (Fig 5). The result in annulus fibrosus perforation it. Periodically agitate the syringe to
safe triangle is bounded by the pedicle and inadvertent discography (Fig 12). resuspend crystals.
superiorly, the exiting nerve medially, In patients with foraminal stenosis or NRB generates diagnostic informa-
and the vertebral body anteriorly (64) lateral disk herniation, target the ven- tion from pain provocation during nee-
(Fig 10). Veins, and sometimes the ra- tral ramus peripheral to the foramen to dle placement, pain provocation during
dicular artery, course through the safe avoid severe pain production during in- injection, immediate pain relief from
triangle, explaining the frequency of jection and failed delivery of medication the anesthetic, or delayed pain relief
vascular cannulation and the “unsafe (Figs 4, 10). from the corticosteroid. In dictated re-
triangle” moniker (108). Curved nee- NRB at S1 requires epidural needle ports, record the provocative response
dles may have advantages over straight placement and poses unique access (pain production during injection) as
needles for evading vessels and navigat- challenges. Both transforaminal and concordant or nonconcordant and the
ing obstructions (eg, hypertrophic facet transosseous techniques are feasible af- immediate analgesic response (pain re-
joints). ter excluding Tarloff cysts and dural ec- duction after injection). Selective NRB
If veins compromise supraneural in- tasia during MR image review. The dor- is intended to yield only diagnostic in-
jection, reposition the needle inferiorly sal S1 foramen is constant in location formation. An anesthetic, not a steroid,
and posteriorly in the foramen. This in- and orientation but variable in caliber. is injected. In selective NRB, the needle
franeural approach targets the Kambin When the foramen is narrow, transfo- tip touches the ventral ramus peripheral
triangle and favors dorsal and caudal raminal navigation can be difficult or to the foramen and provokes radicular
epidural flow (109). The Kambin trian- impossible without meticulous fluoro- pain that is assessed for concordance
gle is bounded by the exiting nerve su- scopic set-up (Fig 9). A curved needle with typical symptoms. The volume of
periorly, the caudal vertebral body in- (5°–10° along the distal centimeter) anesthetic is limited to avoid spread to
feriorly, and the facet joint posteriorly. helps passage through a small angled nontarget nerves.

680 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016

HOW I DO IT: Spinal Injections for Pain Management Palmer

Figure 10 Figure 11

Figure 11:  Oblique fluoroscopic image shows

the set-up for lumbar NRB. During anterposterior
fluoroscopy (images not shown), the detector was
tilted craniocaudally to align the L5 pedicle with
the caudal margin of the transverse process (thin
curved line), to maximize the space between the
L5 transverse process and the sacrum and to
standardize osseous relationships for reproducible
needle placement. The x-ray beam is parallel to the
L5 superior endplate (arrows). The detector was
rotated laterally to open supraneural (S) and infra-
neural (I) routes between the L5 transverse process,
lateral facet border (thick curved line), and iliac
Figure 10:  Transforaminal lumbosacral needle placements. Schematic drawing of the lumbosacral spine wing (arrowheads). Degree of rotation depends on
shows coronal relationships of nerve roots, nerve root ganglia, and postganglionic spinal nerves. In stenotic morphology of facet and iliac wing, desired needle
foramina, extraforaminal needle placement (N1) targets ventral ramus peripherally. In patent foramina, needle placement relative to foramen, and straight versus
placement can be supraneural (N2) in the safe triangle (∗) or infraneural (N3) in the Kambin triangle (cross). curved needle technique. Straight needle technique
At S1, the needle (N4) crosses the posterior S1 foramen and enters the epidural space inferior to the S1 requires direct trajectory and, therefore, greater
pedicle. detector rotation. A curved needle may improve
navigation through narrow spaces and around
Cervical transforaminal injection.— subtraction fluoroscopy, CT guidance, hypertrophic facets.
Cervical transforaminal injection is indi- and nonparticulate corticosteroid ad-
cated in radiculopathy with or without ministration. Others have questioned position the needle more posteriorly
axial neck pain. Injectionists may follow the benefit of cervical transforaminal and peripherally than usual, especially
procedural protocols established by spe- injection with any technique, given the in older individuals who might have tor-
cialty societies, such as the Spine Inter- difficulty associated with visualization tuous arteries. Skip skin anesthesia in
vention Society (or SIS) (100–116). The of small vessels, including the radicular case the external jugular vein underlies
2004 Spine Intervention Society prac- artery (84,121,122). the desired needle entry site.
tice guidelines for cervical NRB recom- To perform fluoroscopy-guided During the fluoroscopic set-up, tar-
mended intraforaminal positioning of cervical NRB with an anterolateral ap- get the foramen posteriorly to increase
the needle tip as deep as the midpoint proach, place the patient in the prone distance from the vertebral artery and
of articular pillars but never deeper than position and turn his or her head away inferiorly to access the epiradicular
a vertical line connecting the uncinate from the side of injection to shift the ca- space of the exiting nerve and to im-
processes (117). After reports of cata- rotid artery from the needle path. Tra- prove the likelihood of intraforaminal
strophic neurologic injuries from cervical jectory modification may be necessary spread of injectate (Fig 6). Whereas
NRBs, some investigators recommended when MR or CT images show an un- lumbar nerves exit the foramen supe-
protocol modifications (55,75,92,118– usually posterior carotid artery or tor- riorly, cervical nerves exit it inferiorly.
120). Potentially safer techniques in- tuous vertebral artery (123). When MR When the needle is positioned too su-
clude extraforaminal needle placement, images are unavailable for procedural periorly and peripherally, injectate may
a lateral or posterior approach, digital planning, reschedule the injection or flow along the more cranial nontarget

Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 681

HOW I DO IT: Spinal Injections for Pain Management Palmer
Figure 12
Figure 12:  Inadvertant diskogram during NRB in a 51-year-old man with left L3 radiculopathy correlating
with L3–4 intraforaminal disk extrusion. (a) Axial T1-weighted MR image at the L3–4 disk level shows a
large left intraforaminal disk extrusion (arrowheads) displacing left L3 nerve (white arrow). Right L3 nerve
(black arrow) location is normal and surrounded by fat. (b) Anteroposterior fluoroscopic image in the prone
position. Curved needles are present on the left side at L3–4 and L4–5. At L3–4, infraneural (retrodiscal)
needle (black arrowhead) punctured the disk annulus, resulting in intradiscal contrast (white arrowheads).
Radicular symptoms were immediately exacerbated due to distension of the herniation sac (arrow). L4–5 =
L4–5 disk space.
Figure 13 nerve (Fig 13). Needle length depends
on neck girth and target level, and it
ranges from 1.5 to 2.5 inches. Remove
any stylet and flush the 25-gauge needle
with contrast material, filling the hub
prior to insertion to obviate gas de-
livery. Direct the needle to the lateral
margin of the articular pillars, switch-
ing between oblique and posteroante-
rior fluoroscopy to check the needle
trajectory and depth.
Document extravascular needle
placement during contrast material
injection with real-time anteroposte-
rior fluoroscopy. Digital subtraction
angiography may improve vessel detec-
Figure 13:  Inadvertent C7 NRB during at-
tion (115). Exit veins by advancing the
tempted C8 NRB in a 44-year-old man with right C8
needle several millimeters. If needle ad-
radiculopathy correlating with C7-T1 intraforaminal
vancement fails to result in vein exit, re-
disk extrusion. Anteroposterior fluoroscopic image
in the supine position shows the needle (arrow) position the needle more caudally along
targets right C8 nerve at C7-T1 foramen. Pain the expected course of the target nerve.
provocation prevented further needle advancement. Short extension tubing (dead space, 0.4
Needle trajectory was satisfactory, but the needle tip mL) helps to avoid hand exposure dur-
terminated peripheral to lateral masses (black lines), ing fluoroscopy and inadvertent needle
distant from the C8 nerve. Injected contrast material motion during syringe exchange. In-
(arrowheads) flowed along nontarget C7 nerve into ject a nonparticulate corticosteroid.
C6–7 foramen between C6 and C7 pedicles. Black In the cervical spine, nonparticulate
 = pedicles from C6-T2. and particulate corticosteroids have
682 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016
Figure 14
Figure 14:  Targeting midline posterior epidural fat in
interlaminar ESI. Midline sagittal reformatted CT image
of the lumbar spine and posterior epidural fat at L2–3
(white ∗) indicates safe zone for needle placement
in ESI. As a general rule, dorsal epidural fat is most
prominent between the bases of spinous processes
(white line between black ∗ at L4 and L5) at the disk
space level (intersection of white and black lines at
L4–5). Needle (N) at L3–4 shows desired tip location in
dorsal epidural fat. Needle trajectory projects cranial to
disk level (black line at L3–4). In normal spines, L5-S1
has the least dorsal epidural fat.
shown comparable short-term effects
(116,124). Immediately after needle
removal, decrease hydrostatic pressure
and the likelihood of hematoma by hav-
ing patients sit upright.
Lumbar interlaminar epidural ste-
roid injection.—Standardized fluoro-
scopic set-up helps to decrease needle
manipulation, radiation dose, and over-
all procedure time (Figs 14, 15). In
younger patients, copious epidural fat
and wide interlaminar spaces facilitate
successful needle placement. Insert
the needle from the side with more se-
vere symptoms, since injectate tends
to spread more to the side of needle
placement. In older patients, degener-
ative curvature, spondylotic deformity,
interlaminar collapse, bony prolifera-
tion, and surgical changes create access
challenges. In patients with rotatory
HOW I DO IT: Spinal Injections for Pain Management Palmer

Figure 15

Figure 15:  Fluoroscopic set-up for L4–5 interlaminar ESI. (a) After tilting the detector crainocaudally during anteroposterior fluoroscopy (not shown) and projecting
L4–5 interlaminar arch (thick curved line) above L4 inferior endplate (arrowheads), laterally rotate the detector 18°–22° to align the apex of the interlaminar arch
between the bases of spinous processes (∗). Interlaminar arch at target level should align with arches above and below (curved thin lines at L3-4 and L5-S1). Degree
of detector rotation depends on scoliosis and rotatory curvature of spine. Align needle (arrow) between bases of spinous processes for midline needle placement in
dorsal epidural fat. (b) Lateral fluoroscopic image shows the needle tip (arrow) enters dorsal epidural space at level of L4–5 disk (straight black line). Needle projects
between L4 and L5 spinous processes (curved black lines). Ventral epidural contrast enhancement (arrowheads) is seen. (c) Anteroposterior fluoroscopic image
shows midline placement of needle (arrow) between L4 and L5 spinous processes (∗). Left paramedian approach was chosen because of asymmetric disk degenera-
tion causing levoconvex curvature and right-sided interlaminar collapse.

scoliosis, approach from the side of the
convex curvature to shorten the needle
throw (Fig 15). Should the needle tip
catch on a lamina or spinous process,
turn the bevel toward bone and twist or
rock the needle gently until it slides off
and advances. Steer away from asym-
metrically thickened ligamentum flavum
and facet synovial cysts. At the level of
hemilaminotomy or hemilaminectomy,
interlaminar access should be contra-
lateral to the surgical bed to avoid peri-
dural adhesions that increase risk of
dural puncture.
A critical juncture approaches as the
needle passes through the ligamentum
flavum. Initially, when force is applied
to the syringe plunger, high pressure
prevents contrast material flow. Sudden
loss of resistance usually means that the
needle has reached the epidural space.
Before making the final decision to
inject the corticosteroid, carefully ob-
serve the contrast material distribution
to exclude intravascular, intrathecal,
retrodural, or intraligamentous spread.
Assume intravascular injection until
proven otherwise (Fig 8). In patients
undergoing lumbar interlaminar ESI,
the lateral view can show both vascular
and intrathecal flow; however, the an-
terposterior view best excludes vessels
when lateral image quality is degraded
due to body habitus. To evade vessels,
advance the needle, redirect it, or rein-
sert it at a different level.
Intrathecal injection shows immedi-
ate dependent layering of contrast ma-
terial in the subarachnoid space. After
dural puncture, ESI can be attempted
at a different level or terminated and
rescheduled to avoid any possibility of
complication due to intrathecal steroid
and anesthetic administration. The
retrodural space (retrodural space of
Okada) can be recognized because it
usually communicates with the inter-
spinous space (125,126). The intraliga-
mentous space is associated with facet
degeneration and ligamentum flavum
delamination (125). In the lateral view,
contrast agent distribution initially
simulates epidurography, but injection
pressure unexpectedly increases after
administration of approximately 1 mL
of the contrast agent. If lateral and an-
teroposterior images enable confirma-
tion of interspinous, interlaminar, or
facet joint opacification, advance the
needle for epidural access.
Epidurographic patterns vary con-
siderably between patients and between
time points in the same patient (127).
Injectate takes the path of least resis-
tance, collecting at the needle tip or
spreading over multiple levels. It may
flow cephalad or caudad, right or left,
circumferentially around the thecal sac
or transforaminally along a nerve root.
The plica mediana dorsalis, a midline
septum that anchors the dural mem-
brane posteriorly, can divide the dorsal
epidural space and restrict injectate flow
unilaterally. Needle repositioning may
be desirable if contrast material spreads
contralateral to the side of symptoms.
Always inject under low pressure.
Spinal stenosis and postoperative

Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 683

HOW I DO IT: Spinal Injections for Pain Management Palmer

peridural scarring can block free flow Figure 16

of the corticosteroid. When injectate
pools locally at a stenotic level or be-
tween stenotic levels, anticipate pain
production. Symptoms are usually tran-
sient when delivering small aliquots and
dissipate after 20–30 seconds. When
larger volumes are injected quickly,
persistent leg symptoms limit delivery
of the full dose and force early termi-
nation of the procedure. At levels of
severe symptomatic spinal stenosis, ESI
may exacerbate symptoms that require
hospitalization for pain control despite
prominent posterior epidural fat.
Transforaminal ESI is an NRB var-
iant procedure intended to be semise-
lective in patients who have unilateral
nonradicular symptoms and nonselec-
tive in patients who lack interlaminar
access due to multilevel laminectomy,
hardware, or bone graft. The goal of
the procedure is diffuse epidural spread
rather than nerve selectivity. Thus, fo-
raminal patency is more important than
the exact level of injection. Stenotic fo-
ramina should be avoided because of
obstructed epidural flow. Although nee-
dle placement and initial flow patterns
can be identical in transforaminal ESI
and NRB, extra diluent disperses the
corticosteroid in transforaminal ESI.
Lumbar and cervical facet joint in-
jection.—In patients with early oste-
oarthritis, facet-related pain reflects
synovitis or capsulitis and responds to
corticosteroid injection. If the steroid
breaks the inflammatory cycle, pain re-
lief lasts longer than the drug lifespan. In
patients with advanced osteoarthritis, ir-
reversible cartilage loss limits treatment
effectiveness. After 6–8 weeks, partic-
ulate preparations will have dissipated
and symptoms will have predictably re-
turned. Hypertrophic facet degeneration
can be managed with periodic injections
or with medial branch radiofrequency
denervation. Attritional degeneration
poses a difficult problem due to the loss
of bone stock, causing spondylolisthesis,
stenosis, and segmental instability. In-
variably, treatment effects do not last as
long, or they become nonexistent.
Direct needles at cartilage inter-
faces or capsular recesses (Figs 7, 16).
Fluoroscopic set-up should account for
Figure 16:  Pars injection via L4-5 facet joint in a 21-year-old woman with low back pain correlating with
L5 spondylolysis. (a) Sagittal reformatted CT image of lumbar spine shows continuity between L4-5 facet
joint (black arrowheads) and L5 pars defect (white arrowhead). (b) Oblique fluoroscopic image in the prone
position shows contrast material filling the needle hub (black arrow) and flowing away from the needle tip
into the superior recess (white arrow) of right L4-5 facet joint. Contrast material spreads inferiorly into the
conjoined space created by L5 pars defect and inferior L4-5 facet joint recess (arrowheads). Axial CT image
(not shown) demonstrated direct access to right L4-5 facet joint at midarticular level approximately 20° off
midline. L4 = L4 pedicle, S1 = S1 pedicle.
articular orientation, curvature, and de- mandible by turning the head oppo-
generation (Fig 1). When osteophytes site the symptomatic side. Place pil-
block the joint space or when listhesis lows under the chest to flex the neck
deforms it, target capsular recesses to and shorten needle distances. At lower
increase the likelihood of intra-articular cervical levels especially, decrease the
needle placement. Periarticular corti- throw by inserting needles perpendicu-
costeroid injection may offer the same lar to the skin. When the needle touch-
therapeutic benefits as intra-articular es bone, angle the detector parallel to
corticosteroid administration (31,128). the joint space and move the needle tip
Document the extra-articular leak of superiorly or inferiorly into the joint
contrast material into a synovial cyst, recess.
intraligamentous cavity, or retrodural
space (125,126,129). At lumbar levels,
synovial cyst rupture is feasible by plac- Conclusion
ing the needle into the facet joint with Spine intervention and pain manage-
fluoroscopic guidance or directly into ment create rewarding opportunities for
the cyst or facet joint with CT guidance radiologists. Radiologists with appro-
(Fig 7). Facet joints communicate with priate training can take responsibility
pars defects; therefore, they can be for treatment decisions and outcomes,
used to deliver a steroid directly into helping patients delay or avoid surgery.
the pseudoarthrosis (130) (Fig 16). These radiologists build patient-physi-
In the cervical spine, both lateral cian relationships, counseling patients
(joint space) and posterior (capsu- with the goal of improving quality of
lar recess) approaches are feasible. life. Radiologists can leverage their ex-
In posterior approaches, displace the pertise in MR image interpretation to

684 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016

HOW I DO IT: Spinal Injections for Pain Management Palmer
correlate imaging findings with clinical
symptoms and establish pain genera-
tors. They can use their knowledge of
fluoroscopic anatomy to target those
pain generators and plan safe and ef-
fective needle placement.
Acknowledgment: The author thanks Susanne
Loomis, MSci, for her expertise in preparation
of the medical illustrations and Eleni Balasalle,
MEd, for her expertise in the preperation of the
supplemental videos.
Disclosures of Conflicts of Interest: W.E.P. dis-
closed no relevant relationships.
References
1.
Evans W. Intrasacral epidural injec-
tion in the treatment of sciatica. Lancet
1930;216(5597):1225–1229.
2. Goebert HW Jr, Jallo SJ, Gardner WJ,
Wasmuth CE, Bitte EM. Sciatica: treat-
ment with epidural injections of pro-
caine and hydrocortisone. Cleve Clin Q
1960;27:191–197.
3. Goebert HW Jr, Jallo SJ, Gardner WJ,
Wasmuth CE. Painful radiculopathy treat-
ed with epidural injections of procaine and
hydrocortisone acetate: results in 113 pa-
tients. Anesth Analg 1961;40:130–134.
4. Coomes EN. A comparison between epi-
dural anaesthesia and bed rest in sciatica.
BMJ 1961;1(5218):20–24.
5. Barry PJ, Kendall PH. Corticosteroid infil-
tration of the extradural space. Ann Phys
Med 1962;6:267–273.
6. Swerdlow M, Sayle-Creer WS. A study
of extradural medication in the relief of
the lumbosciatic syndrome. Anaesthesia
1970;25(3):341–345.
7. White AH, Derby R, Wynne G. Epidural
injections for the diagnosis and treatment
of low-back pain. Spine (Phila Pa 1976)
1980;5(1):78–86.
8. el-Khoury GY, Ehara S, Weinstein JN,
Montgomery WJ, Kathol MH. Epidural
steroid injection: a procedure ideally per-
formed with fluoroscopic control. Radiol-
ogy 1988;168(2):554–557.
9. Renfrew DL, Moore TE, Kathol MH, el-
Khoury GY, Lemke JH, Walker CW. Cor-
rect placement of epidural steroid injec-
tions: fluoroscopic guidance and contrast
administration. AJNR Am J Neuroradiol
1991;12(5):1003–1007.
10. Furman MB, O’Brien EM, Zgleszewski
TM. Incidence of intravascular penetration
in transforaminal lumbosacral epidural
Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 685
steroid injections. Spine (Phila Pa 1976) 23. Friedly J, Chan L, Deyo R. Increases in
2000;25(20):2628–2632. lumbosacral injections in the Medicare
population: 1994 to 2001. Spine (Phila Pa
11. Furman MB, Giovanniello MT, O’Brien
1976) 2007;32(16):1754–1760.
EM. Incidence of intravascular penetra-
tion in transforaminal cervical epidural 24. Manchikanti L, Singh V, Pampati V, Smith
steroid injections. Spine (Phila Pa 1976) HS, Hirsch JA. Analysis of growth of inter-
2003;28(1):21–25. ventional techniques in managing chronic
pain in the Medicare population: a 10-year
12. Smuck M, Fuller BJ, Yoder B, Huerta J.
evaluation from 1997 to 2006. Pain Physi-
Incidence of simultaneous epidural and
cian 2009;12(1):9–34.
vascular injection during lumbosacral
transforaminal epidural injections. Spine J 25. Levinson DR. Medicare payments for facet
2007;7(1):79–82. joint injection services. Department of
Health and Human Services Office of Inspec-
13. Smuck M, Tang CT, Fuller BJ. Incidence
tor General on Medicare Payments for Facet
of simultaneous epidural and vascular
Joint Injection Services. http://oig.hhs.gov/
injection during cervical transforaminal
oei/reports/oei-05-07-00200.pdf. Published
epidural injections. Spine (Phila Pa 1976)
2009;34(21):E751–E755.
September 2008. Accessed January 10,
2015.
14. Jensen MC, Brant-Zawadzki MN, Obu-
26. Carrino JA, Morrison WB, Parker L, Sch-
chowski N, Modic MT, Malkasian D, Ross
weitzer ME, Levin DC, Sunshine JH. Spi-
JS. Magnetic resonance imaging of the
nal injection procedures: volume, provider
lumbar spine in people without back pain.
distribution, and reimbursement in the
N Engl J Med 1994;331(2):69–73.
U.S. medicare population from 1993 to
15. Brant-Zawadzki MN, Dennis SC, Gade GF, 1999. Radiology 2002;225(3):723–729.
Weinstein MP. Low back pain. Radiology
27. Abbott ZI, Nair KV, Allen RR, Akuthota
2000;217(2):321–330.
VR. Utilization characteristics of spinal in-
16. Johnson BA, Schellhas KP, Pollei SR. terventions. Spine J 2012;12(1):35–43.
Epidurography and therapeutic epidural
28. Friedly JL, Comstock BA, Turner JA, et al.
injections: technical considerations and
A randomized trial of epidural glucocorti-
experience with 5334 cases. AJNR Am J
coid injections for spinal stenosis. N Engl J
Neuroradiol 1999;20(4):697–705.
Med 2014;371(1):11–21.
17. el-Khoury GY, Renfrew DL. Percutaneous
29. Carette S, Marcoux S, Truchon R, et al. A
procedures for the diagnosis and treatment
controlled trial of corticosteroid injections
of lower back pain: diskography, facet-joint
into facet joints for chronic low back pain.
injection, and epidural injection. AJR Am J
N Engl J Med 1991;325(14):1002–1007.
Roentgenol 1991;157(4):685–691.
30. Carette S, Leclaire R, Marcoux S, et al.
18. Deyo RA, Mirza SK, Martin BI, Kreuter W,
Epidural corticosteroid injections for sciat-
Goodman DC, Jarvik JG. Trends, major
ica due to herniated nucleus pulposus. N
medical complications, and charges associ-
Engl J Med 1997;336(23):1634–1640.
ated with surgery for lumbar spinal stenosis
in older adults. JAMA 2010;303(13):1259– 31. Lilius G, Laasonen EM, Myllynen P,
1265. Harilainen A, Grönlund G. Lumbar facet
joint syndrome. a randomised clinical trial.
19. Frymoyer JW. Back pain and sciatica. N
J Bone Joint Surg Br 1989;71(4):681–684.
Engl J Med 1988;318(5):291–300.
32. Pinto RZ, Maher CG, Ferreira ML, et al.
20. Manchikanti L, Pampati V, Boswell MV,
Epidural corticosteroid injections in the
Smith HS, Hirsch JA. Analysis of the
management of sciatica: a systematic re-
growth of epidural injections and costs in
view and meta-analysis. Ann Intern Med
the Medicare population: a comparative
2012;157(12):865–877.
evaluation of 1997, 2002, and 2006 data.
Pain Physician 2010;13(3):199–212. 33. Staal JB, de Bie R, de Vet HC, Hildebrandt J,
Nelemans P. Injection therapy for subacute
21. Manchikanti L, Pampati V, Falco FJ,
and chronic low-back pain. Cochrane Da-
Hirsch JA. Assessment of the growth of
tabase Syst Rev 2008;(3):CD001824.
epidural injections in the medicare popu-
lation from 2000 to 2011. Pain Physician 34. Staal JB, Nelemans PJ, de Bie RA. Spinal
2013;16(4):E349–E364. injection therapy for low back pain. JAMA
2013;309(23):2439–2440.
22. Friedly J, Chan L, Deyo R. Geographic var-
iation in epidural steroid injection use in 35. Vad VB, Bhat AL, Lutz GE, Cammisa F.
medicare patients. J Bone Joint Surg Am Transforaminal epidural steroid injections
2008;90(8):1730–1737. in lumbosacral radiculopathy: a prospec-
HOW I DO IT: Spinal Injections for Pain Management Palmer
tive randomized study. Spine (Phila Pa
1976) 2002;27(1):11–16.
36. Watts RW, Silagy CA. A meta-analysis on
the efficacy of epidural corticosteroids in
the treatment of sciatica. Anaesth Inten-
sive Care 1995;23(5):564–569.
37. Koes BW, Scholten RJ, Mens JM, Bouter
LM. Efficacy of epidural steroid injections
for low-back pain and sciatica: a systematic
review of randomized clinical trials. Pain
1995;63(3):279–288.
38. Riew KD, Yin Y, Gilula L, et al. The effect
of nerve-root injections on the need for op-
erative treatment of lumbar radicular pain.
a prospective, randomized, controlled,
double-blind study. J Bone Joint Surg Am
2000;82-A(11):1589–1593.
39. Bresnahan BW, Rundell SD, Dagadakis
MC, et al. A systematic review to assess
comparative effectiveness studies in epi-
dural steroid injections for lumbar spinal
stenosis and to estimate reimbursement
amounts. PM R 2013;5(8):705–714.
40. Bush K, Hillier S. A controlled study of
caudal epidural injections of triamcinolone
plus procaine for the management of in-
tractable sciatica. Spine (Phila Pa 1976)
1991;16(5):572–575.
41. Snoek W, Weber H, Jørgensen B. Double
blind evaluation of extradural methyl pred-
nisolone for herniated lumbar discs. Acta
Orthop Scand 1977;48(6):635–641.
42. Olmarker K, Byröd G, Cornefjord M,
Nordborg C, Rydevik B. Effects of methyl-
prednisolone on nucleus pulposus-induced
nerve root injury. Spine (Phila Pa 1976)
1994;19(16):1803–1808.
43. Olmarker K, Rydevik B. Pathophysiol-
ogy of sciatica. Orthop Clin North Am
1991;22(2):223–234.
56.
44. Olmarker K, Blomquist J, Strömberg J,
Nannmark U, Thomsen P, Rydevik B. In-
flammatogenic properties of nucleus pulpo-
sus. Spine (Phila Pa 1976) 1995;20(6):
665–669.
45. McCarron RF, Wimpee MW, Hudkins PG,
Laros GS. The inflammatory effect of nu-
cleus pulposus. a possible element in the
pathogenesis of low-back pain. Spine (Phi-
la Pa 1976) 1987;12(8):760–764.
46. Saal JS, Franson RC, Dobrow R, Saal JA,
White AH, Goldthwaite N. High levels of
inflammatory phospholipase A2 activity in
lumbar disc herniations. Spine (Phila Pa
1976) 1990;15(7):674–678.
59.
47. Rydevik B, Brown MD, Lundborg G. Patho-
anatomy and pathophysiology of nerve
686 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016
root compression. Spine (Phila Pa 1976)
1984;9(1):7–15.
48. Robinson JP, Brown PB. Medications in low
back pain. Phys Med Rehabil Clin N Am
1991;2(1):97–126.
49. Derby R, Lee SH, Date ES, Lee JH, Lee
CH. Size and aggregation of corticosteroids
used for epidural injections. Pain Med
2008;9(2):227–234.
50. Benzon HT, Chew TL, McCarthy RJ, Ben-
zon HA, Walega DR. Comparison of the
particle sizes of different steroids and the
effect of dilution: a review of the relative
neurotoxicities of the steroids. Anesthesi-
ology 2007;106(2):331–338.
51. Tiso RL, Cutler T, Catania JA, Whalen K.
Adverse central nervous system sequelae
after selective transforaminal block: the
role of corticosteroids. Spine J 2004;4(4):
468–474.
52. Baker R, Dreyfuss P, Mercer S, Bogduk N.
Cervical transforaminal injection of corti-
costeroids into a radicular artery: a possi-
ble mechanism for spinal cord injury. Pain
2003;103(1-2):211–215.
53. Kennedy DJ, Dreyfuss P, Aprill CN, Bog-
duk N. Paraplegia following image-guided
transforaminal lumbar spine epidural ste-
roid injection: two case reports. Pain Med
2009;10(8):1389–1394.
54. Houten JK, Errico TJ. Paraplegia after
lumbosacral nerve root block: report of
three cases. Spine J 2002;2(1):70–75.
55. Rathmell JP, Benzon HT, Dreyfuss P, et al.
Safeguards to prevent neurologic compli-
cations after epidural steroid injections:
consensus opinions from a multidisci-
plinary working group and national or-
ganizations. Anesthesiology 2015;122(5):
974–984.
Scanlon GC, Moeller-Bertram T, Ro-
manowsky SM, Wallace MS. Cervical
transforaminal epidural steroid injections:
more dangerous than we think? Spine (Ph-
ila Pa 1976) 2007;32(11):1249–1256.
57. Windsor RE, Storm S, Sugar R, Nagula
D. Cervical transforaminal injection: re-
view of the literature, complications, and
a suggested technique. Pain Physician
2003;6(4):457–465.
58. Okubadejo GO, Talcott MR, Schmidt RE,
et al. Perils of intravascular methylprednis-
olone injection into the vertebral artery.
an animal study. J Bone Joint Surg Am
2008;90(9):1932–1938.
U.S. Food and Drug Administration.
Drug Safety Communications. FDA Drug
Safety Communication: FDA requires la-
bel changes to warn of rare but serious
neurologic problems after epidural corti-
costeroid injections for pain. http://www.
fda.gov/downloads/Drugs/DrugSafety/
UCM394286.pdf. Published April 23, 2014.
Accessed January 10, 2015.
60. American Society of Anesthesiologists.
ASA Formally Responding to FDA Warn-
ing on Injection of Corticosteroids into
Epidural Space. https://www.asahq.org/
advocacy/fda-and-washington-alerts/
washington-alerts/2014/07/asa-formal-
ly-responding-to-fda-warning-on-injec-
tion-of-corticosteroids-into-epidural-
space?page=5. Published July 14, 2014.
Accessed January 10, 2015.
61. Hancock MJ, Maher CG, Latimer J, et al.
Systematic review of tests to identify the
disc, SIJ or facet joint as the source of
low back pain. Eur Spine J 2007;16(10):
1539–1550.
62. Schaufele MK, Hatch L, Jones W. Inter-
laminar versus transforaminal epidural
injections for the treatment of symptom-
atic lumbar intervertebral disc herniations.
Pain Physician 2006;9(4):361–366.
63. Lee JH, An JH, Lee SH. Comparison of the
effectiveness of interlaminar and bilateral
transforaminal epidural steroid injections
in treatment of patients with lumbosacral
disc herniation and spinal stenosis. Clin J
Pain 2009;25(3):206–210.
64. Pfirrmann CW, Oberholzer PA, Zanetti
M, et al. Selective nerve root blocks for
the treatment of sciatica: evaluation of
injection site and effectiveness—a study
with patients and cadavers. Radiology
2001;221(3):704–711.
65. Bureau NJ, Kaplan PA, Dussault RG. Lum-
bar facet joint synovial cyst: percutane-
ous treatment with steroid injections and
distention—clinical and imaging follow-up
in 12 patients. Radiology 2001;221(1):
179–185.
66. Jackson RP, Jacobs RR, Montesano PX.
1988 Volvo award in clinical sciences. facet
joint injection in low-back pain. a prospec-
tive statistical study. Spine (Phila Pa 1976)
1988;13(9):966–971.
67. Revel ME, Listrat VM, Chevalier XJ, et al.
Facet joint block for low back pain: identi-
fying predictors of a good response. Arch
Phys Med Rehabil 1992;73(9):824–828.
68. Schwarzer AC, Wang SC, O’Driscoll D,
Harrington T, Bogduk N, Laurent R. The
ability of computed tomography to identify
a painful zygapophysial joint in patients
with chronic low back pain. Spine (Phila
Pa 1976) 1995;20(8):907–912.
HOW I DO IT: Spinal Injections for Pain Management Palmer
69. el Barzouhi A, Vleggeert-Lankamp CL,
Lycklama à Nijeholt GJ, et al. Magnetic
resonance imaging in follow-up assessment
of sciatica. N Engl J Med 2013;368(11):
999–1007.
70. Kim KY, Wang MY. Magnetic resonance
image-based morphological predictors of
single photon emission computed tomog-
raphy-positive facet arthropathy in pa-
tients with axial back pain. Neurosurgery
2006;59(1):147–156; discussion 147–156.
71. Helbig T, Lee CK. The lumbar facet
syndrome. Spine (Phila Pa 1976) 1988;
13(1):61–64.
72. McCall IW, Park WM, O’Brien JP. Induced
pain referral from posterior lumbar ele-
ments in normal subjects. Spine (Phila Pa
1976) 1979;4(5):441–446.
73. Cohen SP, Raja SN. Pathogenesis, diagno-
sis, and treatment of lumbar zygapophy-
sial (facet) joint pain. Anesthesiology
2007;106(3):591–614.
74. Gangi A, Dietemann JL, Mortazavi R,
Pfleger D, Kauff C, Roy C. CT-guided inter-
ventional procedures for pain management
in the lumbosacral spine. RadioGraphics
1998;18(3):621–633.
75. Sutter R, Pfirrmann CW, Zanetti M,
Hodler J, Peterson CK. CT-guided cervical
nerve root injections: comparing the im-
mediate post-injection anesthetic-related
effects of the transforaminal injection with
a new indirect technique. Skeletal Radiol
2011;40(12):1603–1608.
76. Timpone VM, Hirsch JA, Gilligan CJ,
Chandra RV. Computed tomography guid-
ance for spinal intervention: basics of tech-
nique, pearls, and avoiding pitfalls. Pain
Physician 2013;16(4):369–377.
77. Koizuka S, Nakajima K, Mieda R. CT-guid-
ed nerve block: a review of the features of
CT fluoroscopic guidance for nerve blocks.
J Anesth 2014;28(1):94–101.
78. Wagner AL. CT fluoroscopic-guided cervi-
cal nerve root blocks. AJNR Am J Neuro-
radiol 2005;26(1):43–44.
79. Wagner AL. Selective lumbar nerve root
blocks with CT fluoroscopic guidance:
technique, results, procedure time, and
radiation dose. AJNR Am J Neuroradiol
2004;25(9):1592–1594.
80. Wagner AL. CT fluoroscopy-guided epidural
injections: technique and results. AJNR Am
J Neuroradiol 2004;25(10):1821–1823.
81. Somayaji HS, Saifuddin A, Casey AT,
Briggs TW. Spinal cord infarction following
therapeutic computed tomography-guided
Radiology: Volume 281: Number 3—December 2016  n  radiology.rsna.org 687
left L2 nerve root injection. Spine (Phila Pa
1976) 2005;30(4):E106–E108.
82. Edlow BL, Wainger BJ, Frosch MP, Copen
WA, Rathmell JP, Rost NS. Posterior cir-
culation stroke after C1-C2 intraarticular
facet steroid injection: evidence for dif-
fuse microvascular injury. Anesthesiology
2010;112(6):1532–1535.
83. Suresh S, Berman J, Connell DA. Cer-
ebellar and brainstem infarction as a
complication of CT-guided transforaminal
cervical nerve root block. Skeletal Radiol
2007;36(5):449–452.
84. Hodler J, Boos N, Schubert M. Must we
discontinue selective cervical nerve root
blocks? report of two cases and review of
the literature. Eur Spine J 2013;22(Suppl
3):S466–S470.
85. Murtagh FR. Computed tomography and
fluoroscopy guided anesthesia and steroid
injection in facet syndrome. Spine (Phila
Pa 1976) 1988;13(6):686–689.
86. Artner J, Cakir B, Reichel H, Lattig F.
Radiation dose reduction in CT-guided
sacroiliac joint injections to levels of
pulsed fluoroscopy: a comparative study
with technical considerations. J Pain Res
2012;5:265–269.
87. Schmid G, Schmitz A, Borchardt D, et
al. Effective dose of CT- and fluoroscopy-
guided perineural/epidural injections of the
lumbar spine: a comparative study. Cardio-
vasc Intervent Radiol 2006;29(1):84–91.
88. Shepherd TM, Hess CP, Chin CT, Gould R,
Dillon WP. Reducing patient radiation dose
during CT-guided procedures: demonstra-
tion in spinal injections for pain. AJNR Am
J Neuroradiol 2011;32(10):1776–1782.
89. Smuck M, Fuller BJ, Chiodo A, et al. Ac-
curacy of intermittent fluoroscopy to detect
100. Sullivan WJ, Willick SE, Chira-Adisai W,
intravascular injection during transforami-
nal epidural injections. Spine (Phila Pa
1976) 2008;33(7):E205–E210.
90. Hoang JK, Apostol MA, Kranz PG, et al.
CT fluoroscopy-assisted cervical transfo-
raminal steroid injection: tips, traps, and
use of contrast material. AJR Am J Roent-
genol 2010;195(4):888–894.
91. Kranz PG, Amrhein TJ, Gray L. Incidence
of inadvertent intravascular injection dur-
ing CT fluoroscopy-guided epidural ste-
roid injections. AJNR Am J Neuroradiol
2015;36(5):1000–1007.
92. Pottecher P, Krausé D, Di Marco L, et al.
Cervical foraminal steroid injections under
CT guidance: retrospective study of in situ
contrast aspects in a serial of 248 cases.
Skeletal Radiol 2015;44(1):1–8.
93. Orth RC, Wallace MJ, Kuo MD; Tech-
nology Assessment Committee of the So-
ciety of Interventional Radiology. C-arm
cone-beam CT: general principles and
technical considerations for use in inter-
ventional radiology. J Vasc Interv Radiol
2008;19(6):814–820.
94. Leschka SC, Babic D, El Shikh S, Woss-
mann C, Schumacher M, Taschner CA.
C-arm cone beam computed tomography
needle path overlay for image-guided pro-
cedures of the spine and pelvis. Neuroradi-
ology 2012;54(3):215–223.
95. Horlocker TT, Wedel DJ, Rowlingson JC,
et al. Regional anesthesia in the patient
receiving antithrombotic or thrombolytic
therapy: American Society of Regional
Anesthesia and Pain Medicine Evidence-
Based Guidelines (Third Edition). Reg
Anesth Pain Med 2010;35(1):64–101.
96. Nam KH, Choi CH, Yang MS, Kang DW.
Spinal epidural hematoma after pain con-
trol procedure. J Korean Neurosurg Soc
2010;48(3):281–284.
97. Bellini M, Barbieri M. Coagulation manage-
ment in epidural steroid injection. Anaes-
thesiol Intensive Ther 2014;46(3):195–199.
98. Horlocker TT, Bajwa ZH, Ashraf Z, et al.
Risk assessment of hemorrhagic complica-
tions associated with nonsteroidal antiin-
flammatory medications in ambulatory pain
clinic patients undergoing epidural steroid
injection. Anesth Analg 2002;95(6):1691–
1697.
99. Shetty SK, Nelson EN, Lawrimore TM,
Palmer WE. Use of gadolinium chelate to
confirm epidural needle placement in pa-
tients with an iodinated contrast reaction.
Skeletal Radiol 2007;36(4):301–307.
et al. Incidence of intravascular uptake in
lumbar spinal injection procedures. Spine
(Phila Pa 1976) 2000;25(4):481–486.
101. Kay J, Findling JW, Raff H. Epidural tri-
amcinolone suppresses the pituitary-adre-
nal axis in human subjects. Anesth Analg
1994;79(3):501–505.
102. Al-Shoha A, Rao DS, Schilling J, Peterson
E, Mandel S. Effect of epidural steroid
injection on bone mineral density and
markers of bone turnover in postmen-
opausal women. Spine (Phila Pa 1976)
2012;37(25):E1567–E1571.
103. Mandel S, Schilling J, Peterson E, Rao DS,
Sanders W. A retrospective analysis of
vertebral body fractures following epidural
steroid injections. J Bone Joint Surg Am
2013;95(11):961–964.
HOW I DO IT: Spinal Injections for Pain Management Palmer
104. Stambough JL, Booth RE Jr, Rothman RH.
Transient hypercorticism after epidural
steroid injection: a case report. J Bone
Joint Surg Am 1984;66(7):1115–1116.
105. Abram SE. Treatment of lumbosacral ra-
diculopathy with epidural steroids. Anes-
thesiology 1999;91(6):1937–1941.
106. MacMahon PJ, Eustace SJ, Kavanagh EC.
Injectable corticosteroid and local anes-
thetic preparations: a review for radiolo-
gists. Radiology 2009;252(3):647–661.
107. Thomas E, Cyteval C, Abiad L, Picot MC,
Taourel P, Blotman F. Efficacy of transfo-
raminal versus interspinous corticosteroid
injection in discal radiculalgia: a prospec-
tive, randomised, double-blind study. Clin
Rheumatol 2003;22(4-5):299–304.
117. Bogduk N, Kentfield CA; International
108. Glaser SE, Shah RV. Root cause analysis
of paraplegia following transforaminal
epidural steroid injections: the ‘unsafe’
triangle. Pain Physician 2010;13(3):237–
244.
109. Park JW, Nam HS, Cho SK, Jung HJ, Lee
118. Hoang JK, Massoglia DP, Apostol MA,
BJ, Park Y. Kambin’s triangle approach of
lumbar transforaminal epidural injection
with spinal stenosis. Ann Rehabil Med
2011;35(6):833–843.
110. Ma DJ, Gilula LA, Riew KD. Complica-
tions of fluoroscopically guided extrafo-
raminal cervical nerve blocks. an analysis
of 1036 injections. J Bone Joint Surg Am
2005;87(5):1025–1030.
111. Gilula LA, Ma D. A cervical nerve block
approach to improve safety. AJR Am J
120. Narouze SN, Vydyanathan A, Kapural L,
Roentgenol 2007;189(3):563–565.
112. Wallace MA, Fukui MB, Williams RL, Ku
A, Baghai P. Complications of cervical se-
lective nerve root blocks performed with
fluoroscopic guidance. AJR Am J Roent-
121. Huntoon MA. Anatomy of the cervical in-
genol 2007;188(5):1218–1221.
113. Kolstad F, Leivseth G, Nygaard OP. Trans-
foraminal steroid injections in the treat-
ment of cervical radiculopathy. a prospec-
tive outcome study. Acta Neurochir (Wien)
2005;147(10):1065–1070; discussion 1070.
114. Schellhas KP, Pollei SR, Johnson BA,
Golden MJ, Eklund JA, Pobiel RS. Selec-
tive cervical nerve root blockade: expe-
688 radiology.rsna.org  n Radiology: Volume 281: Number 3—December 2016
rience with a safe and reliable technique
using an anterolateral approach for nee-
dle placement. AJNR Am J Neuroradiol
2007;28(10):1909–1914.
115. McLean JP, Sigler JD, Plastaras CT, Gar-
van CW, Rittenberg JD. The rate of detec-
tion of intravascular injection in cervical
transforaminal epidural steroid injections
with and without digital subtraction angi-
ography. PM R 2009;1(7):636–642.
116. Lee JW, Park KW, Chung SK, et al. Cer-
vical transforaminal epidural steroid injec-
tion for the management of cervical radicu-
lopathy: a comparative study of particulate
versus non-particulate steroids. Skeletal
Radiol 2009;38(11):1077–1082.
Spine Intervention Society. Cervical trans-
foraminal injection of corticosteroids. In:
Bogduk N, ed. Practice guidelines for spi-
nal diagnostic and treatment procedures.
San Francisco, Calif: International Spine
Intervention Society, 2004; 237–248.
Lascola CD, Eastwood JD, Kranz PG.
CT-guided cervical transforaminal ste-
roid injections: where should the needle
tip be located? AJNR Am J Neuroradiol
2013;34(3):688–692.
119. Wolter T, Mohadjer M, Berlis A, Knoeller
S. Cervical CT-guided, selective nerve
root blocks: improved safety by dor-
sal approach. AJNR Am J Neuroradiol
2009;30(2):336–337.
Sessler DI, Mekhail N. Ultrasound-guided
cervical selective nerve root block: a flu-
oroscopy-controlled feasibility study. Reg
Anesth Pain Med 2009;34(4):343–348.
tervertebral foramina: vulnerable arteries
and ischemic neurologic injuries after
transforaminal epidural injections. Pain
2005;117(1-2):104–111.
122. Engel A, King W, MacVicar J; Standards Di-
vision of the International Spine Intervention
Society. The effectiveness and risks of fluo-
roscopically guided cervical transforaminal
injections of steroids: a systematic review
with comprehensive analysis of the published
data. Pain Med 2014;15(3):386–402.
123. Fitzgerald RT, Bartynski WS, Collins HR.
Vertebral artery position in the setting of
cervical degenerative disease: implica-
tions for selective cervical transforaminal
epidural injections. Interv Neuroradiol
2013;19(4):425–431.
124. Dreyfuss P, Baker R, Bogduk N. Comparative
effectiveness of cervical transforaminal in-
jections with particulate and nonparticulate
corticosteroid preparations for cervical ra-
dicular pain. Pain Med 2006;7(3):237–242.
125. Huang AJ, Palmer WE. Incidence of inad-
vertent intra-articular lumbar facet joint
injection during fluoroscopically guided in-
terlaminar epidural steroid injection. Skel-
etal Radiol 2012;41(2):157–162.
126. Murthy NS, Maus TP, Aprill C. The retro-
dural space of Okada. AJR Am J Roent-
genol 2011;196(6):W784–W789.
127. Whitlock EL, Bridwell KH, Gilula LA.
Influence of needle tip position on injec-
tate spread in 406 interlaminar lumbar
epidural steroid injections. Radiology
2007;243(3):804–811.
128. Lynch MC, Taylor JF. Facet joint injection
for low back pain. a clinical study. J Bone
Joint Surg Br 1986;68(1):138–141.
129. McCormick CC, Taylor JR, Twomey LT.
Facet joint arthrography in lumbar spon-
dylolysis: anatomic basis for spread of con-
trast medium. Radiology 1989;171(1):193–
196.
130. Shipley JA, Beukes CA. The nature of the
spondylolytic defect. demonstration of a
communicating synovial pseudarthrosis in
the pars interarticularis. J Bone Joint Surg
Br 1998;80(4):662–664.