Ultrasound-guided Lumbar Facet Nerve Block: A Sonoanatomic Study of a New Methodologic Approach

To assess the sonographic landmarks and to develop the necessary ultrasound views, a human vertebral column, totally immersed in water, was first examined under high-resolution ultrasound with a 15-MHz linear transducer (Sonos 4500; Philips Medical Systems, Andover, MA). The latter provided highest resolution of the relevant bony structures. The target point for the block, defined as lying on the upper edge of the transverse process and in the groove at the base of the superior articular process, where the medial branch is traversing the upper edge of the transverse process, ventrocranial to the mamilloaccessory ligament, 8was visualized by ultrasound in a cross-axis view and a long-axis view (fig. 1). Then, a 21-gauge, 8-cm needle was placed with its tip at the target point on the transverse process of the third lumbar vertebra and was fixed in this position. The cross- and long-axis ultrasound view with the needle tip at the target were recorded, and the sonographic details were correlated to the normal cross-sectional anatomy by means of cross-sectional preparations derived from an embalmed cadaver (figs. 2 and 3).

In the second part of the cadaver study, to assess the accuracy of needle guidance by ultrasound, these two views were reproduced in a prone-positioned fresh non-embalmed cadaver with four different transducers (2- to 6-MHz curved array, 3- to 11-MHz linear array, 2- to 4-MHz echo scanner, and 15-MHz linear probe; Sonos 4500). The 2- to 6-MHz curved array transducer provided the best results; hence, it was chosen for all further parts of the cadaver study.

Subsequently, the transverse process of the third lumbar vertebra had to be identified in the aforementioned ultrasound long-axis view by counting the transverse processes from the reflex of the sacrum upward. Then, after aligning the transverse process of the third lumbar vertebra with the center beam of the ultrasound sector, the transducer was rotated 90° to depict the characteristic cross-axis view through the articular and the transverse process of the third lumbar vertebra corresponding to the plane described above (fig. 2). Now, under real-time ultrasound guidance (fig. 4), a 21-gauge, 8-cm needle was inserted 6 cm laterally from the midline, on the lateral end of the transducer exactly in the ultrasound plane at an angle of approximately 45° to the skin (fig. 5) and was advanced in a lateral to medial direction until the needle tip reached the target and bony contact was felt. Then, the needle tip was aligned with the center beam of the ultrasound sector, resulting in a slight lateral movement of the transducer toward the needle. Subsequently, the transducer was rotated back to the long-axis view to assure the optimal position of the needle tip at the cranial edge of the transverse process, and if necessary, the needle position was corrected under ultrasound monitoring. Then, back in the cross-axis view, 1 ml saline solution, 0.9%, was injected through the needle, and the spread of the fluid was observed with ultrasound imaging and recorded on videotape. The same procedure was repeated for the fourth and fifth lumbar segments. The cadaver was then carefully dissected to show the actual positions of the three needle tips (fig. 6), and the result was documented.

After institutional review board approval (ethics committee of the University of Vienna, Medical Faculty, General Hospital, Vienna, Austria) and written informed consent, 20 healthy adult volunteers were enrolled. Spinal deformities or pregnancy were criteria for exclusion. Body weight and height were recorded; body mass index and body surface area were calculated. 9The volunteers were placed prone with the hips supported by pillows to compensate for lumbar lordosis. Because this part of the study was noninvasive, no punctures were performed in any volunteer. Ultrasound examinations were performed by one ultrasound-experienced investigator(M. G.) with a 3.5-MHz curved array transducer (Sonos 1500; Hewlett-Packard/Philips, Andover, MA) bilaterally at L3, L4, and L5, as described. Sonographic views were considered feasible for guidance of lumbar facet nerve block when the necessary landmarks (transverse and articular processes) could be delineated in the respective plane. The quality of sonography was rated semiquantitatively as good (clear visibility of the relevant structures), sufficient (impeded visibility, location possible), or insufficient (undisputed location of the relevant structures impossible) in each volunteer. The following sonographic measurements of distances were performed and recorded bilaterally at L3, L4, and L5: vertical distances (depth) from the skin to the target point at the transverse process (d-TP) and to the most superficial part of the articular process (d-AP); lateral distances from the anatomic midline to the target (l-tgTP) and to the end of the transverse process (l-endTP) in the cross-axis view; width of the transverse process (w-TP) and width of the interspaces from the transverse process to the next one above (w-intTP) in the long-axis view, measured at the left-sided L4 level in all volunteers. All depth measurements were computed in the center of the sonograms (fig. 7) to minimize geometric measurement errors. 10All lateral distance measurements were achieved at skin level between the projection points precisely above the structures identified with ultrasound.

To test the clinical feasibility of our new method, a series of 28 ultrasound-guided lumbar facet nerve blocks was performed in five patients with suspected bilateral lumbar facet syndrome, again after local review board approval (ethics committee of the University of Vienna, Medical Faculty, General Hospital) and written informed consent. Coagulopathies or pregnancy were criteria for exclusion. Patients were placed prone with compensation for lumbar lordosis. After sterile skin preparation, subcutaneous local anesthesia was administered with 1.5 ml mepivacaine, 1%, per puncture site. Needle placement exactly followed the protocol described above. After the needles (21 gauge, 8 cm long) had been placed under ultrasound guidance with the 3.5-MHz curved array transducer (Sonos 1500) on one side, standard C-arm fluoroscopy was performed in the posteroanterior and oblique view (fig. 8), and needle position was evaluated according to current guidelines. 11If the needle tip deviated from the ideal position on the upper end of the transverse process close to the superior articular process in the posteroanterior view or from “high on the eye of the Scottie dog” in the oblique view, the divergence was recorded and corrected thereafter under fluoroscopic guidance. Then, 1 ml bupivacaine, 0.25%, was injected slowly per needle, and the needles were removed. Subsequently, the same procedure was performed on the other side.

Statistical analysis was performed with SPSS for Windows 10.0 (SPSS Inc., Chicago, IL). Data were tested for normality by using the Kolmogorov-Smirnov test. Volunteer and patient characteristics are presented as median (range). Distances are expressed as mean ± SD (range) and were analyzed for between-level differences (L3, L4, L5) by means of a one-way analysis of variance with the Bonferroni test post hoc. P  values less than 0.05 were considered to indicate statistical significance.

The sonographic landmarks and the target point could be identified at all three lumbar levels (L3–L5) in the fresh nonembalmed cadaver with the 2- to 6-MHz curved array scanner. Following the described methodology, each level was correctly recognized, and the needles could be traced in real-time from skin puncture to contact with the target point in the cross-axis view without difficulty at each level. To reach the target, the angle of the needle was modified from exactly 45° according to the individual sonogram after skin puncture. The additional long-axis view was essential for the orientation in the craniocaudal direction, which seems not to be sufficiently granted by the cross-axis view alone because we had to slightly reposition the needle tip toward the upper edge of the transverse process in the long-axis view in two of three cases. Injection of 1 ml saline solution, 0.9%, under ultrasound monitoring promptly resulted in a clearly visible spread around the needle tip in all cases. Subsequent dissections of the paravertebral region of the cadaver revealed that all three needles were correctly placed, with the tips located exactly at the predefined target position (fig. 6).

Paravertebral sonography with the 3.5-MHz transducer was feasible in all 20 volunteers (11 men, 9 women; median age, 36 yr [23–67 yr]; height, 176 cm [150–190 cm]; weight, 70 kg [52–90 kg]; body mass index, 23 kg/m²[19–36 kg/m²]; body surface area, 1.8 m²[1.5–2.1 m²]). The sonographic landmarks were delineated with good quality in 19 volunteers and with sufficient quality in one volunteer. This volunteer was obese, with a body mass index of 36 kg/m². In all cases, the relevant sonographic landmarks were depicted.

A total of 240 cross- and long-axis views of the target at the transverse process were obtained to compute ultrasound-derived distance measurements: l-tgTP, 30 ± 3 mm (20–35 mm); l-endTP, 52 ± 5 mm (45–65 mm); w-TP, 14 ± 2 mm (10–16 mm); and w-intTP, 17 ± 2 mm (13–21 mm). All measured data were distributed normally according to the Kolmogorov-Smirnov test. Depths to the target point (d-TP) and to the most superficial part of the articular process (d-AP) at the respective side and level are summarized in table 1. Both increased from L3 to L5 bilaterally, and differences were statistically significant for d-TP between the levels L3 and L5 (right side:P = 0.027; left side:P = 0.005) as shown by a one-way analysis of variance and post hoc  testing (Bonferroni).

A total of 28 ultrasound-guided lumbar facet nerve blocks (8 at L3, 10 at L4, and 10 at L5) were performed by the main investigator (M. G.) in five patients (two women, three men; median age, 51 yr [31–68 yr]; height, 172 cm [163–182 cm]; weight, 78 kg [68–81 kg]). All patients had chronic low back pain for more than 6 months and rated their pain intensity higher than 39 on a visual analog scale (VAS 0–100) before the block procedure. They all showed distinct paravertebral lumbar tenderness but no radicular neurologic symptoms, a negative straight leg–raising test, and no pain increase on Valsalva maneuver. Four out of five felt an aggravation of pain during hyperextension.

All needles could be guided successfully to the right segment by ultrasound. After fluoroscopic control, only 3 of 28 needle positions had to be corrected slightly, the other 25 were placed accurately. All three corrections were within the range of 5 mm to the radiologically defined target point of the classic technique, 11two of them (L3, L5) slightly lateral but still on the dorsal surface of the transverse process and one (L5) a little dorsal on the lateral surface of the superior articular process. Small adaptations of needle direction during ultrasound guidance were common and necessary. However, no multiple puncture attempts after failed positioning occurred. The block procedure in every patient comprised four to six blocks and was finished within at least 40 min including the time for fluoroscopy control. There were no complications, signs of nerve root block, or other neurologic symptoms. Two of five patients were pain free at the evaluation time 30 min later; three had a reduction in pain scores of 50%. All patients could be discharged immediately after the evaluation.

In the current article, we present our new methodology of ultrasound-guided lumbar facet nerve block. This three-part study describes the development of the technique in cadavers, evaluates it noninvasively in volunteers, and finally tests in a small series of patients, whether the technique is usable under clinical circumstances. Unlike ultrasound-guided blocks in other regions, where the nerves are detectable superficially, direct visualization of the small facet nerve was unfeasible at a depth of 5 cm. Nevertheless, our data show that, using this method based on a sonographic cross- and long-axis view, the target point for a lumbar facet nerve block at the level L3–L5 (i.e. , medial branch L2–L4) can be delineated with ultrasound, and a needle can be guided to it. We demonstrated that the placement of this needle can be monitored in real time from skin puncture to the final position at the transverse process with ultrasound and that injection of 1 ml solution results in a striking spread around the needle tip. These features may be considered crucial to ensure specificity and to avoid complications related to needle malpositioning. The essential ultrasound landmarks could be depicted in every case and corresponded well in all three parts of the study. Ultrasound in sufficient quality was feasible even in the most obese volunteer, with a body mass index of 36 kg/m², although constitutional limitations probably exist.

The clinical relevance of the technique is furthermore supported by the fact that in our case series, 25 of 28 ultrasound-guided needles were correctly positioned, according to fluoroscopy control in two planes. The distance of the remaining three needle tips was found to be less than 5 mm away from the target point, making a block failure unlikely. The ultrasound-guided procedure could be finished in a timely manner and was well accepted by all patients.

In addition, the results of our ultrasound measurements are helpful to estimate the distances before performing the block. The vertical projection of the target point at the skin, for example, was found 3 cm laterally to the midline, and the most superficial part of the articulate process at both sides in all investigated levels was located 1.5 cm closer to the skin than the target point. The depth to the target point showed little variability within the segment in our volunteers but was found to increase significantly from L3 (4.5 cm) to L5 (5 cm) at both sides. This increased depth to the target point, together with the presence of artifacts from the iliac crest, leading to inferior ultrasound image quality, might account for the lower success rate at the L5 level in this case series and in other preliminary clinical data. 12 

Facet joint–mediated pain is not uncommon and may be considered the major cause for low back pain in up to 36% of cases. 13Facet joints are often affected by mechanical derangements or degenerative alterations, and reflex muscular spasm or referred pain can easily arise. In 1933, the facet syndrome  was first described and was defined as lumbosacral pain with or without sciatic pain, particularly associated with a twisting or rotary strain of the lumbosacral region. 14There is some evidence for a higher incidence in women. 15The pain can be unilateral or bilateral and is typically enhanced by hyperextension of the lumbar spine or locally applied pressure on the facet joints. Scoring systems have been developed 16; nevertheless, no specific anatomic or radiologic findings have been shown to correlate with the clinical diagnosis of the facet syndrome until now. Consequently, primarily diagnostic facet nerve blocks are considered in many patients presenting with the above symptoms, necessitating at least a fluoroscopy device if not a computed tomography scanner.

Ultrasound guidance is useful in facilitating peripheral and neuraxial blocks 17and offers direct visualization of the target, adjacent structures, and local anesthetic spread. The advantages of ultrasound guidance include but are not limited to increased success, decreased rates of complications, faster onset of blocks, and reduced amount of local anesthetics. 18–20Ultrasound measurements can even result in suggestions to modify established block techniques. 21 

Ultrasound-guided lumbar facet nerve block has not been performed before, and no detailed description of the ultrasound landmarks necessary for this approach exists. Two studies 22,23describing ultrasound-guided posterior lumbar plexus block, however, showed the feasibility of posterior paravertebral sonography. Another group 24described the use of ultrasound for facet joint infiltration of the lumbar spine, but only for the periarticular region and, as they admit, could not ascertain precise intraarticular local anesthetic application without fluoroscopy and contrast media. Nonetheless, either direct intraarticular injections or facet nerve blocks are considered indispensable for treatment of facet joint–mediated pain. 4Studies do not confirm a significant difference in outcome between intraarticular facet joint injections and facet nerve blocks. 25However, there is evidence favoring the use of the latter over the former because it is easier to perform; less traumatic, especially in patients with degenerative joint obliteration 26; and can be followed by radiofrequency procedures when indicated.

The technique for fluoroscopy-guided lumbar facet nerve blocks has improved during the past decade. Guidelines 11stress the importance of performing these procedures using strict techniques to increase sensitivity and specificity based on studies determining their physiologic effectiveness 27and face validity. 8 

Consequently, the findings of our study, first evaluating ultrasound guidance as an alternative, must be reflected on this past experience, and the advantages and disadvantages must be weighed against the standard fluoroscopy procedure. Absolutely no exposure to radiation for the patient and the practitioner is a clear potential advantage of the ultrasound technique, making it principally applicable also during pregnancy. Although the amount of radiation for fluoroscopy-guided blocks may be small and the effect on patient–practitioner mortality and morbidity unclear, elimination of radiation for any procedure is desirable, especially if is not so small, as during computed tomography scan guidance. Moreover, compared with fluoroscopy or computed tomography scanning, ultrasound is only moderately priced and offers more flexibility.

In contrast, it remains to be proved in larger comparison studies that ultrasound can provide the same accuracy of needle placement as the standard technique. Because facet nerve blocks are primarily diagnostic, precise needle placement and small volumes of local anesthetics are necessary to minimize rates of false-positive blocks. Because of that, the target point of the ultrasound technique was defined according to the guidelines for the fluoroscopy technique and was met exactly in 25 of 28 cases in our patients. However, these findings are limited because of the relatively small sample size and are not applicable to an L5 dorsal ramus block, which was not tested in the current study.

Guidelines for the fluoroscopy technique recommend a local anesthetic volume of only 0.5 ml/block preceded by 0.3 ml contrast agent to rule out venous uptake, which occurs in 8% and contributes to false-negative results. 11We used a total volume of 1 ml/block to not miss visualization of solution spread around the needle tip. This might theoretically have caused false-positive results due to aberrant local anesthetic spread (epidural, nerve root, multifidus muscle). The lateral to medial needle direction in ultrasound guidance compared with the suggested lateral to medial and cranial to caudal needle direction in fluoroscopy guidance 8might have in addition promoted aberrant foraminal spread. However, there were no clinical signs of numbness or motor block in the territory of one or more nerve roots, indicative of such a loss of specificity.

Finally, potential patient discomfort due to longer or more painful procedures under ultrasound could be ruled out in our clinical cases. To evaluate patient acceptance and the influence of performer experience in ultrasound techniques on accuracy of and time for needle placement, future controlled studies are needed.

In conclusion, we described the sonoanatomic details necessary for ultrasound-guided lumbar facet nerve block and demonstrated the principal clinical feasibility of this new approach. Real-time ultrasound guidance of the needle can be performed without difficulty after a little practice. According to our data, ultrasound seems to be a promising guidance technique for this common invasive procedure in pain therapy, which should give rise to novel clinical experiences. If subsequent studies confirm the accuracy of ultrasound guidance, this technique could probably grow to replace fluoroscopy or computed tomography scan in routine cases, thus eliminating radiation exposure and simplifying the clinical setting for lumbar facet nerve blocks.

The authors thank Nathalie Frickey, M.D. (Department of Anesthesiology and General Intensive Care, Medical University of Vienna, Vienna, Austria), for editorial assistance.