Content uploaded by Zachary L Mccormick
Author content
All content in this area was uploaded by Zachary L Mccormick on Oct 24, 2017
Content may be subject to copyright.
Review Article
Risks and Benefits of Ceasing or Continuing
Anticoagulant Medication for Image-Guided
Procedures for Spine Pain: A Systematic
Review
Clark C. Smith, MD, MPH,* Byron Schneider, MD,
†
Zachary L. McCormick, MD,
‡
Jatinder Gill, MD,
§
Vivek Loomba, MD,
¶
Andrew J. Engel, MD,
k
Belinda Duszynski,
kj
and Wade King, MMedSc
MMed(Pain),**; on behalf of the Standards Division
of the Spine Intervention Society
*Columbia University College of Physicians and
Surgeons, New York, New York;
†
Vanderbilt
University, Nashville, Tennessee;
‡
University of
California at San Francisco, San Francisco, California;
§
Beth Israel Deaconess Medical Center, Boston,
Massachusetts;
¶
Henry Ford Health System, Detroit,
Michigan;
k
Affordable Pain Management, Chicago,
Illinois;
kj
Spine Intervention Society, Hinsdale, Illinois,
USA; **Manning Pain Management, Mayo Private
Hospital, Taree, New South Wales, Australia
Correspondence to: Clark C. Smith, MD, MPH, 180
Fort Washington Ave. Ste 199, New York, NY, USA.
Tel: 212-305-3535; Fax: 212-342-8472; E-mail:
cs3028@columbia.edu.
Conflicts of interest: None of the authors has any fi-
nancial conflicts of interest to disclose.
Abstract
Objective. To determine the risks of continuing or
ceasing anticoagulant or antiplatelet medications
prior to image-guided procedures for spine pain.
Design. Systematic review of the literature with
comprehensive analysis of the published data.
Interventions. Following a search of the literature
for studies pertaining to spine pain interventions in
patients on anticoagulant medication, seven
reviewers appraised the studies identified and
assessed the quality of evidence presented.
Outcome Measures. Evidence was sought regard-
ing risks associated with either continuing or ceas-
ing anticoagulant and antiplatelet medication in
patients having image-guided interventional spine
procedures. The evidence was evaluated in accord-
ance with the Grades of Recommendation,
Assessment, Development, and Evaluation system.
Results. From a source of 120 potentially relevant
articles, 14 provided applicable evidence.
Procedures involving interlaminar access carry a
nonzero risk of hemorrhagic complications, regard-
less of whether anticoagulants are ceased or
continued. For other procedures, hemorrhagic
complications have not been reported, and case
series indicate that they are safe when performed
in patients who continue anticoagulants. Three
articles reported the adverse effects of ceasing anti-
coagulants, with serious consequences, including
death.
Conclusions. Other than for interlaminar procedures,
the evidence does not support the view that anti-
coagulant and antiplatelet medication must be ceased
before image-guided spine pain procedures.
Meanwhile, the evidence shows that ceasing anticoa-
gulants carries a risk of serious consequences, includ-
ing death. Guidelines on the use of anticoagulants
should reflect these opposing bodies of evidence.
Key Words. anticoagulant; spine; intervention;
epidural; antiplatelet; hematoma
Introduction
Image-guided interventional procedures are used to
diagnose or treat back, neck, and radicular pain.
Patients undergoing these procedures often have
V
C2017 American Academy of Pain Medicine. All rights reserved. For permissions, please e-mail: jour nals.permissions@oup.com 1
Pain Medicine 2017; 0: 1–11
doi: 10.1093/pm/pnx152
comorbidities such as cerebrovascular or cardiovascular
disorders that require the use of antiplatelet or anti-
coagulant medications. For invasive, interventional pain
procedures, these medications may increase the risk of
hemorrhagic complications. The question that arises is
whether anticoagulant medications should be continued
or withheld when various pain procedures are
performed.
Possible hemorrhagic complications of interventional
spine procedures range from a potentially catastrophic
epidural hematoma to minor injection site oozing.
Theoretically, risks of different bleeding complications
should differ among various interventional spine proce-
dures. Although there is a risk of epidural hematoma in
procedures where the epidural space is accessed, there
is no mechanism by which a properly performed extra-
spinal injection such as a sacroiliac (SI) joint injection,
medial branch block, or intra-articular facet joint injec-
tion could cause an epidural hematoma. Other bleeding
complications such as paraspinal hematoma or injection
site oozing may occur in injections performed outside of
the spinal canal.
The 2015 guidelines of the American Society of
Regional Anesthesia (ASRA) explicitly recommend that
anticoagulants be stopped prior to many interventional
pain procedures [1]. While these guidelines provide an
excellent summary of relevant pharmacology, their rec-
ommendations related to interventional spine proce-
dures are not so clearly based on published evidence
related to adverse events.
Some physicians have expressed concerns that stop-
ping anticoagulant medication for interventional spine
procedures may lead to an increased rate of thrombo-
embolic events [2]. Prior studies in general medical
practice have shown that after cessation of therapeutic
anticoagulation, the risk of stroke may be two to three
times greater, and the risk for other major vascular
events is approximately five to six times greater [3,4].
Without evidence-based assessment of the risks and
benefits of ceasing or continuing anticoagulants prior to
image-guided spine procedures, decision-making be-
tween patients and physicians is based on incomplete
information. Therefore, the Spine Intervention Society
(SIS) assembled a task force to collect and assess the
evidence concerning the safety or lack thereof of image-
guided spine pain procedures in patients taking anti-
coagulant medications. This review reports the results of
that exercise.
Methods
A literature search was performed in both PubMed and
EMBASE for articles published January 1948 through
June 2016, using each of the following keyword search
strategies:
1. bleeding risk: (bleeding OR hematoma) AND (injec-
tion OR denervation OR ablation OR neurotomy)
AND spine;
2. spinal hematoma AND injection;
3. cardiovascular risks: (stroke OR myocardial infarction
OR cardiovascular event OR cerebrovascular event
OR thrombosis OR embolism OR blood clot) AND
(injection OR denervation OR ablation OR neurotomy)
AND spine.
When suitable papers were retrieved, their bibliogra-
phies were reviewed for relevant citations that had not
been identified by the database searches.
The titles and abstracts of articles generated by the lit-
erature search were screened by two of the authors (CS
and BD) for prima facie potential relevance or lack
thereof, using the following criteria. The same criteria
were subsequently applied by the investigators who
reviewed the articles selected.
To be accepted for review, articles needed to report a
complication attributable both to an image-guided spine
pain procedure and the use of anticoagulants; to report
the incidence or prevalence of such complications; to
report thrombo-embolic complications in patients who
ceased anticoagulants; or to report the absence of
complications in a series of patients undergoing a spine
pain procedure who were either taking, continuing, or
ceasing anticoagulants during the conduct of that
procedure.
Articles were excluded if they met one of the following
criteria; report of noncatastrophic complications, such
as local bleeding or paraspinal hematoma; report of
thrombo-embolic events not attributable to the use or
cessation of anticoagulants; lack of sufficient information
to allow conclusions to be drawn; lack of clarity regard-
ing whether anticoagulants were continued or ceased;
or lack of clarity regarding the cause of hemorrhagic or
thrombo-embolic complications.
Seven reviewers formally trained in evidence-based
medicine and specializing in spine care assessed the
articles selected for review. Each article was evaluated
to determine if it provided information on the occurrence
of a complication linked to the procedure performed
and if that complication could be attributed to the con-
tinuation or the cessation of anticoagulant or antiplatelet
agents.
Additionally, reviewers determined if articles provided
data to estimate the risk of a complication. Risk was
calculated as the prevalence of a complication in a con-
secutive series of patients, adjusted for sample size by
the 95% confidence interval of a proportion. For articles
reporting a zero prevalence of complications, the
reviewers adopted the upper 95% confidence limit of
that zero proportion as the highest likely risk of the
Smith et al.
2
complication. For this calculation, the method of Wilson
was used as it is deemed appropriate for small magni-
tudes of prevalence [5].
Thereafter, the resultant body of evidence was assessed
using the Grades of Recommendation, Assessment,
Development and Evaluation (GRADE) system to deter-
mine the quality of the evidence [6,7]. In essence, the
GRADE system asks reviewers to evaluate the body of
evidence transparently with consideration not only to
study design, but also to attributes that would
strengthen or weaken confidence in the estimate of ef-
fect. GRADE provides an initial rating of quality based
on the best available evidence that comprises the body
of knowledge, then further requires consideration of
weaknesses (e.g., risk of bias, indirectness) that merit
downgrading, as well as strengths (e.g., large magni-
tude of effect, dose-response gradient) that would justify
upgrading the rating of the quality of the body of evi-
dence. The published data on both the risks of continu-
ing and ceasing anticoagulation prior to interventional
spine procedures were taken into account, and overall
conclusions were drawn in accordance with the GRADE
system.
Results
A total of 2,657 articles were identified by electronic
search of the literature, of which 213 articles were dupli-
cates (identified by both the bleeding and cardiovascular
risk searches). Review of the titles and abstracts of these
2,444 articles identified 54 articles of potential relevance
to bleeding complications and 20 related to cardiovascu-
lar events. Review of the reference lists of the systematic
reviews and primary studies identified by the initial litera-
ture search identified an additional 46 potentially relevant
articles. All authors reviewed these studies.
Most articles failed to satisfy the inclusion criteria and
were subsequently excluded. Others, whose titles
appeared to promise relevance, were excluded for spe-
cific reasons. Three articles were excluded for lack
of sufficient evidence upon which to draw conclusions
[8–11]. Other articles were excluded because the
patients described were not taking any anticoagulants
[12,13]; insufficient information was reported to deter-
mine whether therapeutic anticoagulation was ceased
or continued [14,15]; the cause of the complication was
unclear [16]; the articles did not disclose whether the
patient was on anticoagulants; or the articles did not re-
port for how long medications had been stopped
[17,18]. Reports of bleeding due to clearly improper nee-
dle placement were excluded as well [19,20]. Reports of
thrombo-embolic complications due to the injection of
particulate steroids were also excluded [21–24].
A large observational study that showed no major
bleeding in a cohort of patients undergoing medial
branch blocks was excluded because it was unclear
whether these patients were on therapeutic anticoagula-
tion [25]. Another large study showed no major bleeding
complications after facet joint injections but was likewise
excluded for lack of clarity as to whether patients were
taking therapeutic anticoagulant medications [26].
The remaining accepted literature provided information
in three domains: 1) reports of hemorrhagic complica-
tions in patients undergoing spine pain procedures and
either continuing or ceasing anticoagulants; 2) reports of
the purported safety of spine procedures in anticoagu-
lated patients; and 3) reports of adverse events in
patients who ceased anticoagulants.
Hemorrhagic Complications
No reports of hemorrhagic complications were found at
any segmental level for transforaminal injections, medial
branch blocks, sacral lateral branch blocks, injections
into zygapophysial joints or the SI joint, or radiofre-
quency neurotomy. Hemorrhagic complications attribut-
able to a spine pain procedure were exclusively
reported for interlaminar procedures, being injections of
steroids performed at various segmental levels, and ei-
ther placement or removal of spinal cord stimulator
leads (Table 1).
Three case reports describe catastrophic bleeding as a
complication of cervical interlaminar injection of steroids
(Table 1). It was not clear whether the injection was
image-guided or not in the first case [27], but the injec-
tion was image-guided in the other two cases [28,29].
In the first case, an epidural hematoma caused lasting
paralysis in a patient who continued to take clopidogrel,
diclofenac, and aspirin during the epidural injection [27].
In the other two cases, anticoagulants were ceased in
accordance with guidelines. One patient had stopped
an aspirin-containing product (Fiorinal) seven days prior
to the procedure and developed a subdural hematoma
and lasting paralysis after the procedure [28]. The other
patient had stopped clopidogrel and developed an epi-
dural hematoma but did not suffer lasting neurologic im-
pairment [29].
One case report described an epidural hematoma
resulting in paraplegia after image-guided thoracic inter-
laminar corticosteroid injection [30]. The patient had
stopped warfarin for seven days and was bridged with
enoxaparin 1 mg/kg twice daily, which was discontinued
24 hours prior to the procedure. On the day of the pro-
cedure, the patient’s international normalized ratio (INR)
was 1.00. She remained on aspirin 81 mg/d for the pro-
cedure. The patient continued to have severe neurologic
deficit despite surgical decompression.
Three cases of epidural hematoma have been reported
following image-guided, lumbar interlaminar injection of
steroids (Table 1)[31–33]. In all cases, patients were tak-
ing warfarin, which was stopped days prior to the pro-
cedure. In two of the cases, INR was checked prior to
the procedure and found to be 1.2 and 1.0, respectively
[31,33]. In one case, INR was not checked after warfarin
was held for six days [32]. In two cases, the patient was
Anticoagulants and Image-Guided Spine Pain Procedures
3
Table 1 Reports of hemorrhagic complications during spine pain procedures
Reference Procedure(s)
Bleeding
Complication(s)
Anticoagulation
Status Intervention/Outcome
Cervical
Benzon et al. [27] "Cervical epidural
steroid injection"
Image guidance not
specified
C2-7 epidural
hematoma
Presenting as incom-
plete quadriplegia
Continued on
“diclofenac, clopi-
dogrel, and possibly
aspirin”
Cervical decompres-
sion and hematoma
evacuation
Lower extremities
remained paralyzed
Reitman et al. [28] Image-guided C4-C5
interlaminar ESI
C2-6 subdural
hematoma
Presenting as incom-
plete quadriplegia
Fiorinal (containing
aspirin) stopped for
7d
Cervical
decompression
Partial motor recovery
initially
Postoperative course
complicated by
meningitis, patient
died from cardiopul-
monary arrest post-
operative day 8
Benyamin et al. [29] Image-guided C7-T1
interlaminar ESI
“C3 to upper thoracic”
epidural hematoma
Clopidogrel stopped
for 12 d
Cervical decompres-
sion resulting in
resolution of neuro-
logic symptoms
Thoracic:
Loomba et al. [30] Image-guided T9-10
interlaminar ESI
Thoracic epidural
hematoma
Presenting on postop
day 6 with numb-
ness and weakness
in bilateral lower
extremities
Coumadin stopped for
7 d, bridged with
enoxaparin until
24 h prior to injec-
tion; coagulation
profile normal prior
to procedure
Continued aspirin
81 mg
Thoracic
decompression
No significant reso-
lution of symptoms,
persistent complete
thoracic spinal cord
injury
Lumbar:
Ain et al. [31] Image-guided L4-5
interlaminar ESI
L1-5 epidural
hematoma
Presenting as pain,
urinary retention,
numbness/weak-
ness in lower
extremities
Coumadin held 6 d
prior, enoxaparin
bridge until 24 h be-
fore ESI
INR 1.2 on day of in-
jection, no other
labs reported
Lumbar
decompression
At 1 wk after decom-
pression, reported
to have “mild inter-
mittent low back
pain and numbness
in her toes on the
left”
Xu et al. [32] "Lumbar ESI with loss
of resistance
technique"
Image guidance not
specified
Lumbar epidural
hematoma
Presenting as pain
and bladder
retention
Warfarin stopped 6 d
preprocedure,
placed on enoxa-
parin bridge, with
last dose given 30 h
preprocedure
Peri-procedure co-
agulation labs,
including INR, not
reported
Aspirin stopped 6 d
preprocedure
Lumbar
decompression
Resolution of
symptoms
(continued)
Smith et al.
4
bridged with enoxaparin 1 mg/kg [31,32]. In the third
case, warfarin was held without low–molecular weight
heparin bridging. Two of the patients did not suffer per-
manent neurologic damage [31,32], but the third suffered
permanent bilateral foot drop [33].
Additional evidence pertaining to interlaminar proce-
dures was found in the literature on spinal cord stimula-
tors (Table 1). One patient who was taking aspirin
81 mg per day suffered an epidural hematoma after
placement of a spinal cord stimulator lead, although
with no lasting neurologic consequences [34]. A case
series reported two patients who developed epidural
hematomas after removal of stimulation leads [35]. One
patient continued to take aspirin, including on the morn-
ing of the procedure, whereas the other patient had
stopped aspirin for seven days. The patient who contin-
ued aspirin suffered permanent neurologic injury. A large
study of bleeding complications after percutaneous spi-
nal cord stimulator procedures included 101 procedures
Table 1 Continued
Reference Procedure(s)
Bleeding
Complication(s)
Anticoagulation
Status Intervention/Outcome
Page et al. [33] Image-guided L3-4
interlaminar ESI
Epidural hematoma at
L3-4 radiating to
L4-5
Presenting as foot
drop, urinary reten-
tion, and severe
radicular pain
Coumadin held for 7
d; INR 1.0 on day
of procedure
Lumbar decompres-
sion
2 y postoperation,
there was “little
change in the bilat-
eral foot drop,” but
bowel and bladder
function mostly
recovered
Spinal Cord Stimulator:
Buvanendran et al. [34] Percutaneous spinal
cord stimulator trial
and lead removal
T2-T10 epidural
hematoma
Severe pain in low
back and bilateral
lower extremities
during trial accom-
panied by cessation
of stimulation sen-
sation, prompting
immediate lead re-
moval, followed by
lower extremity
weakness
90 minutes
postremoval
Daily aspirin 81 mg
therapy continued
T5-T10 laminectomy
and hematoma
evacuation 7 h after
symptom onset
Complete resolution
of symptoms
Giberson et al. [35] Removal of percutan-
eous spinal cord
stimulator trial
leads
Patient 1: T5-L2 epi-
dural hematoma
Struggled to sit up
immediately after
lead removal, com-
plained of severe
back pain, followed
quickly by
paraparesis
Patient 2: T8-L3 epi-
dural hematoma
45 minutes after lead
removal, reported
acute lower thoracic
pain and lower ex-
tremity weakness
Patient 1: Excedrin
(containing aspirin)
taken on the morn-
ing of procedure
Patient 2: regular
81 mg aspirin, dis-
continued 11 d be-
fore lead removal
Patient 1: immediate
intravenous steroids,
“laminectomy and
hematoma evac-
uation” 2 d after
presentation
“Permanent weak-
ness in his left leg”
Patient 2: immediate
T8-L1 laminectomy
and hematoma
evacuation
Complete resolution
of symptoms
ESI ¼epidural steroid injection; INR ¼international normalized ratio.
Anticoagulants and Image-Guided Spine Pain Procedures
5
in patients in whom nonsteroidal anti-inflammatory
drugs and aspirin were continued with no bleeding
complications [36]. The 95% confidence interval of this
zero prevalence is 0% to 3.7% due to small sample
size, limiting a meaningful estimate of zero or low risk.
Estimates of Safety
Several large observational studies [37–39] reported no
cases of major bleeding in various numbers of patients
who continued anticoagulants while undergoing various
spine pain procedures (Table 2).
For radiofrequency neurotomy and various cervical pro-
cedures, the numbers of patients monitored were too
small to provide a meaningful estimate of zero risk. For
other procedures, however, more substantial samples of
patients were monitored.
Studies have attempted to estimate the prevalence of
bleeding complications among patients who underwent
lumbar transforaminal injections while continuing anti-
coagulant and antiplatelet medications. One study found
no bleeding complications in 90 cases in which anti-
coagulant and antiplatelet drugs were continued [37].
Another larger study reported no complications in 955
patients who continued anticoagulants and none in 663
patients who continued antiplatelet medications (Table
2)[38]. These figures indicate that the 95% confidence
interval for hemorrhagic complications is 0.0% to 0.4%
when continuing anticoagulants and 0.0% to 0.6%
when continuing antiplatelet medications.
There are no reports of bleeding complications among
patients who continued therapeutic anticoagulation for
lumbar medial branch blocks. Endres et al. (2016) [38]
reported no complications in 1,142 patients who contin-
ued anticoagulants during lumbar medial branch blocks
or in 925 who continued antiplatelet medications. When
calculating 95% confidence intervals, these figures indi-
cate a risk of hemorrhagic complications of less than
0.3% for continuing anticoagulants and less than 0.4%
for continuing antiplatelet medications.
Among patients who underwent SI joint blocks, there
were no complications in 174 who continued anticoagu-
lants or in 85 who continued antiplatelet medications
[38]. The upper limits of the 95% confidence intervals
for the respective risks of hemorrhagic complications,
such as paraspinal hematoma, were 2.2% and 4.3%.
For lumbar facet intra-articular injections, one study
found no complications among 58 injections where
therapeutic anticoagulant and antiplatelet medications
were continued [37]. Another study encountered no
bleeding complications in 1,109 patients who continued
antiplatelet medications during intra-articular injections
of the lumbar zygapophysial joints (Table 2)[39]. This
constitutes a risk for epidural hematoma of less than
0.3%. Table 2 includes a list of all anticoagulants and
antiplatelet agents used in the above studies.
Medical Complications
The literature contains two case reports of medical
complications in patients who ceased anticoagulants
prior to a spine pain procedure (Table 3)[40,41]. One
patient developed a pulmonary embolism after stopping
warfarin in preparation for a spinal cord stimulator trial
[40]; the other patient suffered a middle cerebral artery
stroke after discontinuing warfarin for a lumbar epidural
injection of steroids [41].
One large retrospective study monitored 2,218 cases in
which anti-thrombotic therapy was stopped (Table 3)
[39]. No major cardiovascular complications were
encountered. In a large, prospective study, anticoagu-
lants were ceased in 1,626 patients [38]. Nine patients
suffered thrombo-embolic complications (Table 3). One
died from a stroke. Another died from a myocardial in-
farction. Five suffered nonfatal strokes, one suffered a
myocardial infarction, and one suffered a pulmonary
embolism. The prevalence of these complications was
nine in 1,626 (0.4%; 95% confidence interval [CI] ¼
0.2–0.7%). However, all events occurred in patients
who ceased warfarin. The same study demonstrated no
thromboembolic complications among 701 patients who
stopped antiplatelet therapy [38].
GRADE Evaluation
For procedures for which there is no evidence either for
risk or for safety, no conclusion about the quality of evi-
dence is applicable. For other procedures, the quality of
evidence differs according to the type of evidence
available.
The evidence concerning safety of interlaminar proce-
dures is low in quality according to the GRADE system
because it is based solely on case reports. This level
could be upgraded if future studies reported the preva-
lence of complications in prospective case series.
For lumbar transforaminal injections, lumbar medial
branch blocks, and lumbar intra-articular injections, the
GRADE quality of evidence is moderate as it is based
on large observational studies. For SI joint injections, the
quality of evidence is perhaps best regarded as low be-
cause although provided by an observational study, the
sample size is small. Larger sample sizes could upgrade
the quality of evidence to moderate.
Discussion
The published literature clearly shows that interlaminar
procedures carry a risk of hemorrhagic complications.
The magnitude of that risk is not known because no
case series has provided a prevalence estimate, but it is
manifestly nonzero. Indeed, the majority of case reports
of hemorrhagic complications occurred in patients who
ceased anticoagulant and antiplatelet medications
according to the 2015 ASRA guidelines. This predomin-
ance may reflect the fact that during routine practice
Smith et al.
6
Table 2 Case series of interventional spine procedures performed without ceasing anticoagulation
Reference Procedure(s) Bleeding Complications Anticoagulation Status
Moeschler et al. [36] 421 patients, 642 percu-
taneous SCS proce-
dures (SCS trial,
revision, or
implantation)
No major bleeding
complications
101 procedures on patients who had
taken aspirin or NSAIDs within 7 d of
procedure
Goodman et al. [37] 197/4,253 procedures
were performed on
patients taking thera-
peutic anticoagulant
and antiplatelet
medications
One patient not taking
therapeutic or antipla-
telet medication suf-
fered spinal epidural
hematoma, which
resolved with nonsurgi-
cal conservative care
Anticoagulant* and antiplatelet
†
medica-
tions were continued:
90 LTFESI
11 lumbar intradiscal injection
4 CMBB
58 lumbar facet injection
3 sympathetic paravertebral injection
3 cervical RFN
23 lumbar RFN
5 SIJ injection
*Warfarin, dabigatran, apixaban,
rivaroxaban
†
Clopidogrel, cilostazol, prasugrel,
ticagrelor
Endres et al. [38] 1,383 patients, 7,062
pain procedures
No major bleeding
complications
On anticoagulants* at time of procedure:
955 LTFESI
1,142 LMBB
174 SIJ injection
35 RFN
26 ILESI
On antiplatelets
†
at time of procedure:
633 LTFESI
925 LMBB
85 SIJ injection
22 RFN
15 ILESI
‡
*Warfarin, rivaroxaban, dabigatran, apix-
aban, enoxaparin
†
Aspirin/dypyridanole, clopidogrel, cilos-
tazol, ticagrelor, prasugrel
‡
Subtherapeutic only
Manchikanti et al. [39] ESI: 2,664/10,261
patients on antithrom-
bosis medication
Facet: 2,068/7,482 on
antithrombosis
medication
Other injections: 183/546
on antithrombosis
medications
No major bleeding com-
plications that resulted
in neurologic injury
On warfarin at time of procedure:
10 ESI
7 facet procedures
2 “other”
On antiplatelet* at time of procedure:
1,455 ESI
1,109 facet procedures
78 “other”
*Aspirin, clopidogrel, aspirin
CMBB ¼cervical medial branch block; cont ¼continued; dc ¼discontinued; ESI ¼epidural steroid injection; facet ¼facet joint in-
jection; ILESI ¼interlaminar epidural steroid injection; LMBB ¼lumbar medial branch block; LTFESI ¼lumbar transforaminal epi-
dural steroid injections; RFN¼radiofrequency neurotomy; SIJ ¼sacroiliac joint injection.
Anticoagulants and Image-Guided Spine Pain Procedures
7
anticoagulants are held in the vast majority of patients
undergoing procedures. The occurrence of bleeding
complications was not limited to any particular antiplate-
let or anticoagulant. While different antiplatelet agents
are associated with varying degrees of platelet inhibition,
there is insufficient evidence to distinguish the risk of
bleeding complications among these agents. Although
the risk of hemorrhagic complications following interlam-
inar procedures might be small, some reported compli-
cations are severe or catastrophic. Therefore, the
published evidence supports safety concerns during
interlaminar procedures, regardless of whether anti-
coagulant and antiplatelet agents are continued or
ceased.
For several other procedures, published evidence is ei-
ther lacking or minimal in magnitude. There were no
case reports that met inclusion criteria of hemorrhagic
complications for cervical transforaminal injections, cer-
vical medial branch blocks, or radiofrequency
Table 3 Cardiovascular and thrombo-embolic complications attributable to discontinuing
anticoagulation or antiplatelet therapy for pain procedures
Reference Technique
Cardiovascular
Complication(s) Anticoagulation Status
Endres et al. [38] 1,383 patients, 7,062 pain
procedures
9/1,626 complications when
anticoagulants stopped
0/701 complications when
antiplatelet stopped
Complications: 2 patients
died, 5 suffered strokes, 1
pulmonary embolism, 1
myocardial infarction
Anticoagulants held* at time of
procedure:
505 LTFESI
458 LMBB
41 SIJ injection
309 RFN
180 ILESI
Antiplatelets
†
held at time of
procedure:
250 LTFESI
178 LMBB
6 SIJ injection
195 RFN
72 ILESI
*Warfarin, rivaroxaban, dabiga-
tran, apixaban, enoxaparin
†
Aspirin/dypyridanole, clopidog-
rel, cilostazol, ticagrelor,
prasugrel
Manchikanti et al. [39] ESI: 2,664/10,261 patients
on antithrombosis
medication
Facet injection: 2,068/7,482
on antithrombosis
medication
Other injections: 183/546 on
antithrombosis
medications
No major cardiovascular
complications reported in
2,218 cases in which
antithrombosis was
stopped
48-hour follow-up
Warfarin held for procedure:
182 ESI
238 facet procedures
22 “other”
Antiplatelet* held procedure:
981 ESI
714 facet procedures
81 “other”
*Aspirin, clopidogrel, aspirin,
and others
Kumar et al. [40] Dorsal Column Stimulator
Trial
Pulmonary embolism
No lasting complications
Warfarin discontinued
Linn et al. [41] L5-S1 epidural steroid
injection
Right middle cerebral artery
infarction; persistent left
hemiparesis, neglect, and
dysarthria
Warfarin discontinued for 9 d
preprocedure
cont ¼continued; dc ¼discontinued; ESI ¼epidural steroid injection; ILESI ¼interlaminar epidural steroid injection; inj ¼injection;
MBB ¼medial branch block; RFN ¼radiofrequency neurotomy; SIJ ¼sacroiliac joint injection; TFESI ¼transforaminal epidural
steroid injections.
Smith et al.
8
neurotomy at any segmental level; conversely, no case
series has provided compelling data that show that
these procedures are safe in patients who continue anti-
coagulants. Such case series are needed before these
procedures might be regarded as safe.
In contrast, the published evidence indicates that lum-
bar transforaminal injections, lumbar intra-articular injec-
tions, and lumbar medial branch blocks are safe and
that SI joint injections might be safe, regardless of con-
tinuation of therapeutic anticoagulation or antiplatelet
medication. For these procedures, there are no case
reports of hemorrhagic complications. Large case series
allow for the calculation of the 95% confidence interval for
the true prevalence of bleeding complications. For lumbar
medial branch blocks, the risk is less than 0.3% in patients
continuing anticoagulants and less than 0.4% in patients
continuing antiplatelet medications. For lumbar transforami-
nal injections, the respective risks are less than 0.4% and
less than 0.6%. For lumbar intra-articular injections, the
risk is less than 0.2% in patients taking antiplatelet medi-
cations. For SI injections, the risks are less than 2.2% for
patients taking anticoagulants and less than 4.3% for
patients taking antiplatelet medications.
Antecedent studies have established a definition of
safety for spine procedures. A study concluded that
lumbar epidural steroid injections are safe in patients
taking anti-inflammatory drugs, on the grounds that zero
hemorrhagic complications were encountered in 383
consecutive patients, for a risk of less than 0.96% [42].
One large study conducted by eminent anesthesiolo-
gists found no bleeding complications in 386 patients
for a risk of 1.1%. The resulting risks of bleeding com-
plications of 0.96% and 1.1% were considered safe by
the authors [42,43].
According to these standards, lumbar medial branch
blocks with a risk of 0.3% or 0.4% and lumbar intra-
articular injections with a risk of 0.3% are clearly safe.
So too are lumbar transforaminal injections, with a risk
of either 0.4% in patients continuing anticoagulants or
0.6% in patients taking antiplatelet agents. The risk of
hemorrhagic complications for SI injections (2.2–4.3%)
falls short of the threshold for safety (1.0%) but might
be demonstrated in a larger case series in the future.
In contrast to the safety of the aforementioned proce-
dures, the literature records complications from ceasing
anticoagulants for spine pain procedures, as currently
recommended by guidelines [1]. Although small (0.4%;
0.2–0.7%), the risk of vascular and cerebrovascular
complications is not negligible, and the nature of the
complications is severe and potentially life-threatening.
Moreover, in patients taking warfarin, the risk of vascular
and cerebrovascular medical complications (0.6%;
0.32–1.15%) is almost significantly greater statistically
than the risk of hemorrhagic complications in like
patients undergoing lumbar medial branch blocks (0.00–
0.34%), lumbar intra-articular injections (0.00–0.35%), or
lumbar transforaminal injections (0.00–0.38%). These
figures warn that guidelines concerning the cessation of
anticoagulants need to balance the risk of hemorrhagic
complications against a potentially greater risk of med-
ical complications.
With respect to vascular and cerebrovascular complica-
tions, this focused review found no vascular or cerebro-
vascular complications pertaining to holding antiplatelet
medications. However, there is a large body of literature
that points to the real and high risk of interrupting antipla-
telet therapy, especially for stent thrombosis [44,45]. This
review only found vascular and cerebrovascular compli-
cations in patients who discontinued warfarin. This finding
may indicate that stopping warfarin may constitute a par-
ticular risk of vascular and cerebrovascular complications.
This may be due to factors relating to the agent itself or
the conditions for which it is prescribed.
This review was undertaken expressly to identify the lit-
erature pertinent to anticoagulation and spine pain pro-
cedures. It was not designed to formulate guidelines or
to discuss conjectures about the risks of complications
in procedures about which there is no evidence.
However, the evidence gathered serves to inform the
directions that future guidelines might take.
For many procedures, there is no published evidence ei-
ther of risk or of safety. Recommendations for these
procedures would be theoretical until evidence becomes
available. Interlaminar procedures carry a risk of hemor-
rhagic complications. In contrast, the evidence does not
warrant attributing any significant risk to lumbar transfor-
aminal injections, lumbar intra-articular injections, or
lumbar medial branch blocks in anticoagulated patients.
Meanwhile, ever present is the risk of serious medical
complications if anticoagulant and antiplatelet medica-
tions are ceased for any spine pain procedure.
In conclusion, the evidence does not support the view
that anticoagulant and antiplatelet medications must be
ceased before image-guided spine pain procedures that
do not involve interlaminar access. Meanwhile, the evi-
dence shows that ceasing anticoagulants carries a risk
of serious consequences, including death. Guidelines
on the use of anticoagulants should reflect these oppos-
ing bodies of evidence.
Acknowledgments
The authors wish to thank the other members of the Spine
Intervention Society’s Standards Division: Professor
Nikolai Bogduk and Drs Anil Sharma, Milan Stojanovic,
and Yakov Vorobeychik, who read and offered comments
on the final draft.
References
1 Narouze S, Benzon HT, Provenzano DA, et al.
Interventional spine and pain procedures in patients
Anticoagulants and Image-Guided Spine Pain Procedures
9
on antiplatelet and anticoagulant medications:
Guidelines from the American Society of Regional
Anesthesia and Pain Medicine, the European
Society of Regional Anaesthesia and Pain Therapy,
the American Academy of Pain Medicine, the
International Neuromodulation Society, the North
American Neuromodulation Society, and the World
Institute of Pain. Reg Anesth Pain Med 2015;40:
182–212.
2 Furman MB, Plastaras CT, Popescu A, et al. Should
antiplatelet medications be held before cervical epi-
dural injections? PMR 2014;6:442–50.
3 Wijn W, Kohl P, Danchin N. Guidelines on myocar-
dial revascularization. The task force on myocardial
revascularization of the European Society of
Cardiology (ESC) and the European Association for
Cardio-Thoracic Surgery (EACTS). Eur Heart J 2010;
31:2501–55.
4 Collet JP, Himbet F, Steg PG. Myocardial infarction
after aspirin cessation in stable coronary artery dis-
ease patients. Int J Cardiol 2000;76:257–8.
5 www.epitools.ausvet.om.au.
6 Balshem H, Helfand M, Schunemann HJ, et al.
GRADE guidelines: 3. Rating the quality of evidence.
J Clin Epidemiol 2011;64:401–6.
7 Guyatt GH, Oxman AD, Kunz R, et al. Going from
evidence to recommendations. BMJ 2008;336:
1049–51.
8 Ghaly RF. Recovery after high-dose methylpredniso-
lone and delayed evacuation: A case of spinal epi-
dural hematoma. J Neurosurg Anesthesiol 2001;13:
323–8.
9 Nam KH, Choi CH, Yang MS, Kang DW. Spinal epi-
dural hematoma after pain control procedure. J
Korean Neurosurg Soc 2010;48:281–4.
10 Schreiber AL, McDonald BP, Kia F, Fried GW.
Cervical epidural steroid injections and spinal cord
injuries. Spine J 2016;16(10):1163–6.
11 Shanthanna H, Park J. Acute epidural haematoma
following epidural steroid injection in a patient with
spinal stenosis. Anaesthesia 2011;66:837–9.
12 Lee JY, Nassr A, Ponnappan RK. Epidural hema-
toma causing paraplegia after a fluoroscopically
guided cervical nerve-root injection. A case report. J
Bone Joint Surg 2007;89A:2037–9.
13 Desai MJ, Dua S. Perineural hematoma following
lumbar transforaminal steroid injection causing
acute-on-chronic lumbar radiculopathy: A case re-
port. Pain Practice 2014;14:271–7.
14 Tan LA, Kasliwal MK, Deutsch H. Rapidly progres-
sive quadriparesis secondary to cervical epidural
hematoma—a rare complication of epidural injec-
tion. Br J Neurosurg 2015;29:280.
15 Williams KN, Jackowski A, Evans PJ. Epidural
haematoma requiring surgical decompression fol-
lowing repeated cervical epidural steroid injections
for chronic pain. Pain 1990;42:197–9.
16 Stoll A, Sanchez M. Epidural hematoma after epi-
dural block: Implications for its use in pain manage-
ment. Surg Neurol 2002;57:235–40.
17 LaBan MM, Kasturi G, Wang IM. Epidural cortico-
steroid injections precipitating epidural hematomas
with spinal paresis. Am J Phys Med Rehab 2007;86:
166–7.
18 Pobiel RS, Schellhas KP, Eklund JA, et al. Selective
cervical nerve root blockade: Prospective study of
immediate and longer term complications. AJNR
2009;30:507–11.
19 Adiguzel E, Kesikburun S, Yasar E, Taskaynatan MA.
Major perirectal hematoma complicating sacroiliac
joint injection. J Musculoskel Pain 2014;22:99–101.
20 Kim M, Park KS. Intracranial chronic subdural
hematoma presenting with intractable headache
after cervical epidural steroid injection. J Korean
Neurosurg Soc 2015;58:144–6.
21 Edlow BL, Wainger BJ, Frosch MP, et al. Posterior
circulation stroke after C1-C2 intraarticular facet
steroid injection: Evidence for diffuse microvascular
injury. Anesthesiology 2010;112:1532–5.
22 Engel A, King W, Macvicar J. The effectiveness and
risks of fluoroscopically guided cervical transforami-
nal injections of steroids: A systematic review with
comprehensive analysis of the published data. Pain
Med 2014;15:386–402.
23 Rozin L, Rozin R, Koehler SA, et al. Death during
transforaminal epidural steroid nerve root block (C7)
due to perforation of the left vertebral artery. Am J
Forensic Med Path 2003;24:351–5.
24 Scanlon GC, Moeller-Bertram T, Romanowsky SM,
Wallace MS. Cervical transforaminal epidural steroid
injections: More dangerous than we think? Spine
2007;32:1249–56.
25 Manchikanti L, Malla Y, Wargo BW, Cash KA,
Pampati V, Fellows B. Complications of
Smith et al.
10
fluoroscopically directed facet joint nerve blocks: A
prospective evaluation of 7,500 episodes with
43,000 nerve blocks. Pain Physician 2012;15:
E143–150.
26 Plastaras C, McCormick Z, Macron D, et al.
Adverse events associated with fluoroscopically
guided zygapophyseal joint injections. Pain
Physician 2014;17:297–304.
27 Benzon HT, Wong HY, Siddiqui T, Ondra S. Caution
in performing epidural injections in patients on several
antiplatelet drugs. Anesthesiology 1999;91:1558–9.
28 Reitman CA, Watters W 3rd. Subdural hematoma
after cervical epidural steroid injection. Spine 2002;
27:E174–6.
29 Benyamin RM, Vallejo R, Wang V, et al. Acute epi-
dural hematoma formation in cervical spine after
interlaminar epidural steroid injection despite discon-
tinuation of Clopidogrel. Reg Anesthes Pain Med
2016;41:398–401.
30 Loomba V, Kaveeshvar H, Dwivedi S. Paraplegia fol-
lowing thoracic epidural steroid injection. A A Case
Rep 2016;7(5):118–21.
31 Ain RJ, Vance MB. Epidural hematoma after epidural
steroid injection in a patient withholding enoxaparin
per guidelines. Anesthesiology 2005;102:701–3.
32 Xu R, Bydon M, Gokaslan ZL, et al. Epidural steroid
injection resulting in epidural hematoma in a patient
despite strict adherence to anticoagulation guide-
lines: Case report. J Neurosurg Spine 2009;11:
358–64.
33 Page J, Moisi M, Oskouian RJ. Lumbar epidural
hematoma following interlaminar fluoroscopically
guided epidural steroid injection. Reg Anesth Pain
Med 2016;41:402–4.
34 Buvanendran A, Young AC. Spinal epidural hema-
toma after spinal cord stimulator trial lead placement
in a patient taking aspirin. Reg Anesth Pain Med
2014;39:70–2.
35 Giberson CE, Barbosa J, Brooks ES, et al. Epidural
hematomas after removal of percutaneous spinal
cord stimulator trial leads: Two case reports. Reg
Anesth Pain Med 2014;39:73–7.
36 Moeschler SM, Warner NS, Lamer TJ, et al.
Bleeding complications in patients undergoing per-
cutaneous spinal cord stimulator trials and implanta-
tions. Pain Med 2016;XXX(X):XXX–XX.
37 Goodman BS, House LM, Vallabhaneni S, et al.
Anticoagulant and antiplatelet management for spi-
nal procedures: A prospective, descriptive study.
Pain Med 2016;XXX(X):XXX–XX.
38 Endres S, Shufelt A, Bogduk N. The risks of con-
tinuing or discontinuing anticoagulants for patients
undergoing common interventional pain procedures.
Pain Med 2016;XXX(X):XXX–XX.
39 Manchikanti L, Malla Y, Wargo BW, et al. A pro-
spective evaluation of bleeding risk of interventional
techniques in chronic pain. Pain Physician 2011;14:
317–29.
40 Kumar V, Turakhia PB, Wunnava MS. Pulmonary
embolism during Dorsal Column Stimulator (DCS)
Trial. Reg Anesth Pain Med 2013;38: (abstract only).
41 Linn AJ, Desilva C, Peeters-Asdourian C.
Thromboembolic stroke: A rare complication associ-
ated with peri-procedural management of an epi-
dural steroid injection. Pain Physician 2009;12:
159–62.
42 Horlocker TT, Bajwa ZH, Ashraf Z, et al. Risk as-
sessment of hemorrhagic complications associated
with nonsteroidal antiinflammatory medications in
ambulatory pain clinic patients undergoing epidural
steroid injection. Anesth Analg 2002;95:1691–7.
43 Horlocker TT, Wedel DJ, Schroeder DR, et al.
Preoperative antiplatelet therapy does not increase
the risk of spinal hematoma associated with regional
anesthesia. Anesth Analg 1995;80:303–9.
44 V
aclav
ık J, T
aborsk
y M. Antiplatelet therapy in the
perioperative period. Eur J Intern Med 2011;22:
26–31.
45 American Society of Anesthesiologists Committee
on Standards and Practice Parameters. Practice
alert for the perioperative management of patients
with coronary artery stents: A report by the
American Society of Anesthesiologists Committee
on Standards and Practice Parameters.
Anesthesiology 2009;110:22–3.
Anticoagulants and Image-Guided Spine Pain Procedures
11
