Abstract

Background Context. Medial branch blocks may have unrecognized vascular uptake potentially resulting in false- negative results.

Purpose. To determine the rate of unintended vascular injection of contrast medium during medial branch blocks (MBB) with digital subtraction (DS) technology in the context of negative vascular uptake as determined by live fluoroscopy.

Study Design/Setting. Prospective Study in an academic medical center.

Patient Sample. 344 consecutive MBBs in 80 subjects.

Outcome Measures. The presence of vascular flow as determined by live fluoroscopy and DS technology.

Methods. Unintended vascular injection of contrast medium was determined on 344 consecutive MBBs in 84 subjects, first using live fluoroscopy followed by DS. If live fluoroscopy initially detected vascular uptake, the needle was repositioned until no vascular flow was detected. Once no vascular uptake was confirmed by live fluoroscopy, a contrast medium was then injected while being visualized with DS to again assess the presence or absence of vascular flow undetected by live fluoroscopy.

Results. Live fluoroscopy revealed inadvertent vascular uptake in 38 of the 344 blocks [11% (95% CI 8.0–15%)]. DS uncovered an additional 27 of the 344 blocks [7.8% (95% CI 5.3–11.4%)] with evidence of vascular uptake that were not detected with conventional live fluoroscopy.

Conclusion. DS enhances the ability to detect inadvertent vascular flow during medial branch blocks. This study demonstrates that standard live fluoroscopy can miss a small percentage of cases with unintentional vascular uptake during MBB when compared with DS and may contribute to occasional false-negative responses.

Introduction

Zygapophysial joints (z-joints) are a known source of spine pain with prevalence rates between 15–45% of patients suffering from low back pain [ 1 , 2 ]. The medial branches of the primary dorsal rami of the spinal nerves innervates the z-joints, with two medial branches innervating each z-joint [ 3 ]. Radiofrequency neurotomy (RF) is commonly used for the treatment of chronic z-joint pain that has been refractory to other conservative treatments [ 4 ]. Successful RF requires both optimal technique and patient selection. Medial branch blocks (MBBs) are a validated diagnostic technique to identify the z-joint as a pain generator [ 5 , 6 ]. Successful pain relief following an MBB has been shown to correlate with successful RF outcomes [ 7 ].

MBBs are associated with occasional inadvertent intravascular injection, which may be a potential source of false-negative response to blocks [ 8 ]. Several techniques exist to determine intravascular injection. Needle aspiration has a low sensitivity of only 34.1% for detecting inadvertent vascular entry when compared with live fluoroscopy [ 9 ]. Spot radiography after contrast medium injection has revealed a 3.5% incidence of inadvertent vascular uptake during lumbar medial branch blocks [ 10 ]. This rate increases to 6.1–8.5% when using live fluoroscopy to detect vascular uptake [ 5 , 11 , 12 ].

Digital subtraction (DS) technology, sometimes referred to a Digital Subtraction Angiography or DSA, is a technique that is known to enhance the detection of inadvertent intravascular injections during other spine procedures [ 13 , 14 ]. DS has never been studied in the context of MBB. The purpose of this study was to determine the rate of inadvertent vascular injection during MBB utilizing DS.

Methods

This study was Health Insurance Portability and Accountability Act Compliant, and institutional review board approved. Data were prospectively collected from a single institution on 356 consecutive medical branch blocks of the spine in 84 subjects from August 2011 through April 2012. Subjects were included if they presented for a first medial branch block, but were not included during repeat injections to avoid duplication of subjects. Injections were preformed applying techniques outlined in the International Spine Intervention Society guidelines [ 15 ]. All procedures were performed by a single physician, Board-certified in Physical Medicine and Rehabilitation, with over 5 years’ experience, with or without a fellow trainee. All procedures utilized either a 22- or a 25-gauge quincke point spinal needle, with lengths of 3.5 or 5 inches. Needles were placed under intermittent fluoroscopic guidance with the final position confirmed by both an anterior-posterior (AP) and lateral fluoroscopic image. Next, the needle stylet was removed and the presence or absence of blood in the needle hub was noted. Active aspiration was not done due to previous studies showing its lack of sensitivity. Regardless of flash results, a small amount (typically 0.2–0.4 mL) of radio-opaque contrast medium (Omnipaque M240, GE Healthcare, Princeton, NJ) was injected through 6-inch microbore extension tubing utilizing live fluoroscopy in the AP view. The treating physician thus recorded the presence or absence of vascular flow for live fluoroscopy. If vascular flow was noted, the needle was repositioned and the contrast medium injection was repeated until there was no vascular flow detected using live fluoroscopy. Thereafter, for each injection with no vascular uptake detected under live fluoroscopy, another 0.2–0.4 mL aliquot of contrast medium was injected through the extension tubing and observed using live DS. The treating physician again noted the presence or absence of vascular flow. Figure 1 and Figure 2 depict the difference in contrast flow visualization with fluoroscopy and with digital subtraction angiography.

Figure 1

Live fluoroscopy. Figure 1 shows a right-sided medial branch block targeting the right L3 and L4 medial branches, and the L5 dorsal ramus to anesthetize the right L4/L5 and L5/S1 zygapophysial joints. This is a spot view taken at the end of a live fluoroscopy run for the L3 medial branch under the AP view.

Figure 1

Live fluoroscopy. Figure 1 shows a right-sided medial branch block targeting the right L3 and L4 medial branches, and the L5 dorsal ramus to anesthetize the right L4/L5 and L5/S1 zygapophysial joints. This is a spot view taken at the end of a live fluoroscopy run for the L3 medial branch under the AP view.

Figure 2

Digital subtraction technology. Figure 2 was taken immediately following Figure 1 during the same procedure on the same patient. It is a spot view taken at the end of a live fluoroscopy run with digital subtraction utilized for the L3 medial branch under the AP view. Venous flow heading superiorly and inferiorly is clearly demonstrated that was not readily visualized with live fluoroscopy.

Figure 2

Digital subtraction technology. Figure 2 was taken immediately following Figure 1 during the same procedure on the same patient. It is a spot view taken at the end of a live fluoroscopy run with digital subtraction utilized for the L3 medial branch under the AP view. Venous flow heading superiorly and inferiorly is clearly demonstrated that was not readily visualized with live fluoroscopy.

Proportions were used to determine the incidence of positive test results for blood in the hub and for vascular uptake using live fluoroscopy. For the latter, only the results of the initial injection of contrast medium were used. Sensitivity and specificity analysis was performed on all injections. For blood in the hub, the initial run of live fluoroscopy was considered the gold standard as in several instances the needle was moved prior to the use of DS. For live fluoroscopy, DS was considered the gold standard. Statistical analyses were performed using Microsoft Excel (Microsoft Corp., Redmond, WA).

Results

There were 344 medial branch blocks performed in 80 patients. Table 1 provides baseline demographics and Table 2 displays results based on spinal level. Blood was observed in the needle hub in 23 of the 344 cases (6.7% [95% CI 4.4–10%]). Compared with subsequent live fluoroscopy, the sensitivity of blood in the needle hub to detect vascular uptake was only 15%, with 82% accuracy. The rate of vascular detection increased under live fluoroscopy to 38 of the 344 cases (11% [95% CI 8.0–15%]) with a sensitivity of 63% and an accuracy of 94% when compared with DS. Following negative results under live fluoroscopy, digital subtraction angiography detected inadvertent vascular uptake in an additional 27 of the 344 cases (7.8% [95% CI 5.3–11.4%]). Of the 27 blocks with a positive vascular flow on DS, 22/27 (81.5% [95% CI 63.3–91.2%]) did not involve repositioning due to vascular flow on live fluoroscopy.

Table 1

Demographics

Sex 101 male (28.4%) 
255 female (71.6%) 
Age Average 60.4 years 
Range (21–80) 
Sex 101 male (28.4%) 
255 female (71.6%) 
Age Average 60.4 years 
Range (21–80) 
Table 2

Vascular rates based on level of injection

Level L5 L4 L3 L2 L1 Total 
n = subjects n = 99 n = 122 n = 95 N = 19 N = 9 n = 344 
Flash positive N = 7 N = 8 N = 7 N = 1 N = 0 N = 23 
7.1% 6.6% 7.4% 5.3% 0% 6.7% 
95% CI 95% CI 95% CI 95% CI 95% CI 95% CI 
(3.5–13.9%) (3.4–12.4%) (3.6–14.4%) (0.9–24.6%) (0–29.9%) (4.4–10%) 
Live Fluoroscopy Positive N = 11 N = 16 N = 10 N = 0 N = 1 N = 38 
11.1% 13.1% 10.5% 0% 11.1% 11.0% 
95% CI 95% CI 95% CI 95% CI 95% CI 95% CI 
(6.3–18.8%) (8.2–20.2%) (5.8–18.3%) (0–16.8%) (1.9–43.5%) (8.0–15%) 
DSA Positive N = 7 N = 8 N = 8 N = 1 N = 3 N = 27 
8.1% 6.6% 8.4% 5.3% 33.3% 7.8% 
95% CI 95% CI 95% CI 95% CI 95% CI 95% CI 
(3.5–13.8%) (3.4–12.4%) (4.3–15.7%) (0.9–24.6%) (12.6–64.6%) (5.3–11.4%) 
Level L5 L4 L3 L2 L1 Total 
n = subjects n = 99 n = 122 n = 95 N = 19 N = 9 n = 344 
Flash positive N = 7 N = 8 N = 7 N = 1 N = 0 N = 23 
7.1% 6.6% 7.4% 5.3% 0% 6.7% 
95% CI 95% CI 95% CI 95% CI 95% CI 95% CI 
(3.5–13.9%) (3.4–12.4%) (3.6–14.4%) (0.9–24.6%) (0–29.9%) (4.4–10%) 
Live Fluoroscopy Positive N = 11 N = 16 N = 10 N = 0 N = 1 N = 38 
11.1% 13.1% 10.5% 0% 11.1% 11.0% 
95% CI 95% CI 95% CI 95% CI 95% CI 95% CI 
(6.3–18.8%) (8.2–20.2%) (5.8–18.3%) (0–16.8%) (1.9–43.5%) (8.0–15%) 
DSA Positive N = 7 N = 8 N = 8 N = 1 N = 3 N = 27 
8.1% 6.6% 8.4% 5.3% 33.3% 7.8% 
95% CI 95% CI 95% CI 95% CI 95% CI 95% CI 
(3.5–13.8%) (3.4–12.4%) (4.3–15.7%) (0.9–24.6%) (12.6–64.6%) (5.3–11.4%) 

Neither needle length nor gauge resulted in a difference in vascular detection rates by any measure and all results had overlapping 95% confidence intervals ( Table 3 ). The fluoroscopy time averaged 39.4 seconds per patient, or 9.33 sec/MBB, including the entire procedure: site marking, needle advancement, live fluoroscopy, and DS. A sample of convenience from the same physician revealed the average fluoroscopy time for an MBB without DS was 16.4 seconds per patient, or 4.3 second per MBB.

Table 3

Needle characteristics

 Flash N = 23/344 Live N = 38/344 DSA N = 27/344 
Needle Length 

5 inch

n = 101

 
N = 7 N = 6 N = 6 
6.9% 95% CI (3.4–13.6%) 5.9% 95% CI (2.8–12.4%) 5.9% 95% CI (2.8–12.4%) 

3.5 inch

n = 254

 
N = 16 N = 32 N = 21 
6.3% 95% CI (3.9–9.9%) [12.6% 95% CI 9.1–17.3%)] 8.2% 95% CI (5.5–12.3%) 
Needle Gauge 

22g

n = 158

 
N = 12 N = 12 N = 9 
7.6% 95% CI 4.4–12.8%) 7.6% 95% CI 4.4–12.8%) 5.7% 95% CI 3.0–10.5%) 

25g

n = 198

 
N = 11 N = 26 N = 18 
5.6% 95% CI 3.1–9.7%) 13.1% 95% CI 9.1–18.5%) 9.1% 95% CI 5.8–14.2%) 
 Flash N = 23/344 Live N = 38/344 DSA N = 27/344 
Needle Length 

5 inch

n = 101

 
N = 7 N = 6 N = 6 
6.9% 95% CI (3.4–13.6%) 5.9% 95% CI (2.8–12.4%) 5.9% 95% CI (2.8–12.4%) 

3.5 inch

n = 254

 
N = 16 N = 32 N = 21 
6.3% 95% CI (3.9–9.9%) [12.6% 95% CI 9.1–17.3%)] 8.2% 95% CI (5.5–12.3%) 
Needle Gauge 

22g

n = 158

 
N = 12 N = 12 N = 9 
7.6% 95% CI 4.4–12.8%) 7.6% 95% CI 4.4–12.8%) 5.7% 95% CI 3.0–10.5%) 

25g

n = 198

 
N = 11 N = 26 N = 18 
5.6% 95% CI 3.1–9.7%) 13.1% 95% CI 9.1–18.5%) 9.1% 95% CI 5.8–14.2%) 

Discussion

Our study contributes new information regarding potential sources of false-negative results in lumbar MBBs. Specifically, this study demonstrated that standard live fluoroscopy missed 7.8% (95% CI 5.3–11.4%) of cases with inadvertent vascular uptake during MBB when compared with DS. While false-negative MBB results do not affect the success rates of RF treatment, they may result in withholding treatment from some who could otherwise benefit from RF neurotomy [ 11 ]. MBBs are the only validated method to predict response to RF neurotomy and select patients for treatment [ 6 , 15 , 16 ]. Prior research efforts have improved the outcomes from medical branch RF neurotomy by increasing the accuracy of diagnostic MBBs, but this research has focused primarily on minimizing false-positive results as opposed to exposing potential causes of false-negative results [ 17 ].

Improving the detection of unintentional vascular injections is important because it helps ensure the appropriate patients are offered RF neurotomy for long-term treatment of their back pain. Kaplan et al. [ 5 ]. reported that when venous uptake occurs (despite subsequent lack of venous uptake with needle repositioning), it carries a 50% risk for false-negative results. Inadvertent intravascular injection detection rates with needle aspiration, spot radiography after contrast medium injection, and real-time fluoroscopy during contrast medium injection have all been reported in other types of spine injections [ 13 ]. In a study by Lee et al. [ 11 ], the sensitivity of the aspiration test to predict intravascular contrast medium uptake was 34.1%. They speculate that aspiration test precision was potentially limited by the low pressures venous system and vessel collapse due to the negative pressures generated by aspiration. In this same study, spot radiographs taken immediately after contrast medium injection showed only a moderate increase in sensitivity at 59.1% to detect vascular uptake observed by live fluoroscopy [ 11 ]. This, and several other studies comparing intravascular injection detection rates, led to suggestions that fluoroscopy-guidance and confirmation of contrast medium spread under live fluoroscopy should be the standard of care in spine injections [ 12 , 18 , 19 ].

While it is good practice, the ability of real-time fluoroscopy to detect intravascular injection may have some limitations. Digital subtraction angiography was shown to increase the rate of vascular flow detection in one study of cervical transforaminal injections [ 13 ]. Likewise, the current study shows that DS enhances the detection of inadvertent intravascular injection. It is, however, not without drawbacks. Recent studies have demonstrated that the use of DS results in a 2.3–4.2 fold increase in radiation dose during lumbar transforaminal epidural steroid injections [ 20 ]. Relative to a convenience group comparison, this study shows that DS is associated with a near doubling of the fluoroscopy time during MBB, which alters the risk/benefit ratio. The potential morbidity associated with inadvertent vascular injection during lumbar transforaminal injection is more serious than what is possible during MBB, so the risk/benefit ratio for DS differs between these procedures. Thus, readers should understand that the results of this study do not suggest routine use of DS for MBBs in clinical practice. Further research is necessary to determine that issue. However, this study does provide new insight into a potential source of false-negative responses to MBB.

The high rates of venous uptake in this study reveal another drawback of uncontrolled single MBB, or strict adherence to a 100% relief threshold to define positive response from MBB, as the degree of relief may be attenuated by undetected venous uptake. In one study, Cohen demonstrated 52% of patients having ≥50% pain relief at 6 months when selected by MBB with ≥50 but <80% relief [ 21 ]. He found only 56% of subjects had ≥50% pain relief at 6 months when selected with ≥80% relief from MBB, which was not a statically significant difference. Unfortunately, percentages of subjects with 80–100% relief were not reported, which may have produced different results. In the Cohen study, however, there was no advantage to utilizing 50% or 80% relief in predicting RF neurotomy outcomes. As there is a concern of not wanting to eliminate false-negative responders, the exact degree of pain relief may be less important than a reproducible relief with dual blocks. A cost analysis by Derby et al. [ 22 ], showed the total cost per patient was lowest with a dual block paradigm. Studies that have utilized dual blocks collectively have twice as many patients with durable pain relief at one year [ 7 , 23 ].

There are several limitations to this study. First, due to practice referral patterns, the vast majority of the cases referred were done in the lumbar spine, and we have only reported the results of the lumbar MBBs. Another limitation is that the venous flow detected under live fluoroscopy was not confirmed with DS. We assumed that when vascular flow was appreciated under live fluoroscopy that it would also be evident using DS. This, however, has not been validated and may affect the outcomes, although we think it is unlikely. It is also a possibility that the DS preferentially showed vascular uptake in those initially positive under live fluoroscopy. Specifically it is possible the initial injection of contrast medium and subsequent repositioning may have diminished the ability to detect subsequent venous flow with live fluoroscopy. However, the majority of DS positive subjects 22/27 (81.5% [95% CI 63.3–91.2%]) did not have any repositioning prior to utilization of DS, thus decreasing the likelihood of this potential bias. Also the inter-rater reliability of intravascular detection has not been validated and this study relied on one experienced attending to determine the presence or absence of vascular flow. However, the rates of venous flow observed were similar to the previously published manuscripts on this topic. The inter-rater reliability of live contrast medium flow patterns and the effects of anticoagulation medications on vascular flow during MBBs are currently being studied by the authors.

Conclusions

The rate of vascular uptake during medial branch blocks detected with live fluoroscopy was 10.7% (95% CI 7.5–13.9%), but digital subtraction enhancement revealed an additional 7.8% (95%CI 5.3–11.4%) of subjects with vascular uptake. This may be a source of false-negative responses, or incomplete pain relief from a diagnostic medial branch block. Due to the increase in fluoroscopy time and related radiation exposure, this technique is not currently recommended for routine clinical use. Physicians should consider whether or not the additional radiation is necessary to avoid a potential false-negative result.

Funding sources: This research was not supported by any industry funds.

References

1
Manejias
EM
Hu
J
Tatli
Y
Lutz
GE.
Lumbar zygapophysial joint radiofrequency denervation: A long-term clinical outcome study
.
Hss J
 
2008
;
4
:
180
7
.
2
DePalma
MJ
Ketchum
JM
Saullo
T.
What is the source of chronic low back pain and does age play a role?
Pain Med
 
2011
;
12
:
224
33
.
3
Bogduk
N
Long
DM.
The anatomy of the so-called ‘articular nerves’ and their relationship to facet denervation in the treatment of low-back pain
.
J. Neurosurg
 
1979
;
51
:
172
7
.
4
Smuck
M
Crisostomo
RA
Trivedi
K
Agrawal
D.
Success of initial and repeated medial branch neurotomy for zygapophysial joint pain: A systematic review
.
Pm R
 
2012
;
4
:
686
92
.
5
Kaplan
M
Dreyfuss
P
Halbrook
B
Bogduk
N.
The ability of lumbar medial branch blocks to anesthetize the zygapophysial joint. A physiologic challenge
.
Spine
 
1998
;
23
:
1847
52
.
6
Bogduk
N.
Evidence-informed management of chronic low back pain with facet injections and radiofrequency neurotomy
.
Spine J
 
2008
;
8
:
56
64
.
7
Macvicar
J
Borowczyk
JM
Macvicar
AM
Loughnan
BM
Bogduk
N.
Lumbar medial branch radiofrequency neurotomy in New Zealand
.
Pain Med
 
2013
;
14
(
5
):
639
45
.
8
Gofeld
M
Faclier
G.
Radiofrequency denervation of the lumbar zygapophysial joints–targeting the best practice
.
Pain Med
 
2008
;
9
:
204
11
.
9
Furman
MB
O’Brien
EM
Zgleszewski
TM.
Incidence of intravascular penetration in transforaminal lumbosacral epidural steroid injections
.
Spine
 
2000
;
25
:
2628
32
.
10
Verrills
P
Mitchell
B
Vivian
D
, et al.   .
The incidence of intravascular penetration in medial branch blocks: Cervical, thoracic, and lumbar spines
.
Spine
 
2008
;
33
:
E174
7
.
11
Lee
CJ
Kim
YC
Shin
JH
, et al.   .
Intravascular injection in lumbar medial branch block: A prospective evaluation of 1433 injections
.
Anesth. Analg
 
2008
;
106
:
1274
8
.
12
Sullivan
WJ
Willick
SE
Chira-Adisai
W
, et al.   .
Incidence of intravascular uptake in lumbar spinal injection procedures
.
Spine
 
2000
;
25
:
481
6
.
13
McLean
JP
Sigler
JD
Plastaras
CT
Garvan
CW
Rittenberg
JD.
The rate of detection of intravascular injection in cervical transforaminal epidural steroid injections with and without digital subtraction angiography
.
PM R
 
2009
;
1
:
636
42
.
14
Jasper
JF.
Role of digital subtraction fluoroscopic imaging in detecting intravascular injections
.
Pain Physician
 
2003
;
6
:
369
72
.
15
Bogduk
N
.
International Spine Intervention Society & Standards Committee
.
Practice Guidelines for Spinal Diagnostic and Treatment Procedures
  . San
Francisco, CA
:
International Spine Intervention Society
,
2004
.
16
Bogduk
N.
On the rational use of diagnostic blocks for spinal pain
.
Neurosurg Quarter
 
2009
;
19
:
88
100
.
17
Lord
SM
Barnsley
L
Wallis
BJ
McDonald
GJ
Bogduk
N.
Percutaneous radio-frequency neurotomy for chronic cervical zygapophyseal-joint pain
.
N Engl J Med
 
1996
;
335
:
1721
6
.
18
Stretanski
MF
Chopko
B.
Unintentional vascular uptake in fluoroscopically guided, contrast-confirmed spinal injections: A 1-yr clinical experience and discussion of findings
.
Am J Phys Med Rehabil
 
2005
;
84
:
30
5
.
19
Smuck
M
Fuller
BJ
Chiodo
A
, et al.   .
Accuracy of intermittent fluoroscopy to detect intravascular injection during transforaminal epidural injections
.
Spine
 
2008
;
33
:
E205.
20
Maus
T
Schueler
BA
Leng
S
, et al.   .
Radiation dose incurred in the exclusion of vascular filling in transforaminal epidural steroid injections: Fluoroscopy, digital subtraction angiography, and CT/fluoroscopy
.
Pain Med
 
2014
;
15
:
1328
33
.
21
Cohen
SP
Stojanovic
MP
Crooks
M
, et al.   .
Lumbar zygapophysial (facet) joint radiofrequency denervation success as a function of pain relief during diagnostic medial branch blocks: A multicenter analysis
.
Spine J
 
2008
;
8
:
498
504
.
22
Derby
R
Melnik
I
Lee
JE
Lee
SH.
Cost comparisons of various diagnostic medial branch block protocols and medial branch neurotomy in a private practice setting
.
Pain Med
 
2013
;
14
(
3
):
378
91
.
23
Bogduk
N
Dreyfuss
P
Govind
J.
A narrative review of lumbar medial branch neurotomy for the treatment of back pain
.
Pain Med
 
2009
;
10
:
1035
45
.