Abstract
To define the source and the course of the articular branches to the midthoracic zygapophysial (“z”) joints.
Cadaveric dissection.
The Gross Anatomy Laboratory of the Duke University School of Medicine.
Ten human cadaveric thoraces.
Gross and stereoscopic dissection of dorsal rami T4-T8 was performed bilaterally on 10 adult embalmed cadavers. The medial and lateral branches were traced to their origins from the dorsal rami, and the course of the articular nerves was documented through digital photography. Radio-opaque wire (20 gauge) was applied to the nerves. Fluoroscopic images were obtained to delineate their radiographic course with respect to osseous landmarks.
Forty-eight inferior articular branches were identified. Three (6.3%) originated from the medial branch and 44 (91.7%) from the dorsal ramus. One was indeterminate. Fifty-one superior articular branches were identified. Eight (15.7%) originated from the medial branch and 43 (84.3%) from the dorsal ramus. In 12% of cases (6/50), there was side-to-side asymmetry in the origins of the articular branches. Nerves were commonly suspended in the intertransverse space. The articular branches contacted an osseous structure in only 39% of cases. As previously reported, a “descending branch” was not identified in any specimen.
Articular branches to the T4-T8 z-joints have substantial inter- and intraspecimen variability of origin. They typically arise from the dorsal ramus rather than the medial branch and frequently do not contact any osseous structure to allow percutaneous needle placement.
Introduction
The spinal zygapophysial (“z”) joints are paired, diarthrodial, synovial joints [1,2]. The z-joints are recognized sources of pain in the cervical and lumbar spine [3–10]. However, diagnosis of z-joint pain has proven confounding. There are no clinical findings, such as patterns of tenderness or aggravation by particular movements, that are diagnostic of lumbar z-joint pain [11,12]. Additionally, computed tomography (CT) has been unable to reliably diagnose lumbar z-joint pain [13]. Anatomical studies have empirically established that the medial branches of cervical and lumbar dorsal rami innervate the cervical and lumbar z-joints, respectively [14,15]. Anesthetization of these structures in a procedure known as “medial branch block,” has become the standard technique of diagnosing cervical and lumbar z-joint pain [16–20].
The thoracic z-joints possess the anatomical substrate for pain transmission [1]. Further, thoracic spine pain can be as disabling as cervical or lumbar spinal pain [21]. In parallel with the cervical and lumbar spine, the diagnosis and treatment of thoracic z-joint pain has focused on identifying the course of the thoracic medial branches and their location with respect to clinically relevant osseous target points, as exemplified by the detailed dissections conducted by Chua and Bogduk [22]. This study documents that the T1-3 and T9-10 medial branches can be found on a relatively consistent basis on the superolateral corner of the transverse process of the subjacent vertebrae (i.e., T2-4 and T10-11, respectively). This consistent relationship to osseous structures is crucial for providing appropriate target points for percutaneous access and subsequent anesthetization.
Despite the important knowledge gained from this study, the resulting technique of interrogating the thoracic z-joints may have some limitations. First, the study by Chua and Bogduk [22] also found that the medial branches of the middle thoracic spine, T4-T8, are commonly found in the soft tissue between the vertebral transverse processes, known as the intertransverse space. The lack of a known bony landmark to serve as an appropriate target has limited the development of a consistent fluoroscopic technique for medial branch block at these levels, though it does not preclude the use of CT guidance. Second, the concept of using medial branch block for diagnostic interrogation of the thoracic z-joints has been called into question by subsequent dissections [23]. In their gross and stereomicroscopic dissection of 120 pairs of thoracic spinal nerves from 10 cadavers, Ishizuka et al. [23] found no instance of articular branches to the thoracic z-joints originating from the medial branches. Instead, this dissection identified the articular branches to the thoracic z-joints as deriving from the first branch of the thoracic dorsal rami, which they termed the “descending branch” [23]. The “descending branch” provided articular branches in nearly half of the cases, with remaining cases having no direct connection to the thoracic z-joints. To further address the anatomic basis for anesthetizing the mid-thoracic z-joints, we conducted this study to clarify the source and course of the articular branches to the T4-T8 z-joints.
Methods
The embalmed cadaveric thoraces of five adult males (aged 64–84 years) and five adult females (aged 53–73 years) were obtained through the Anatomical Gifts Program, Duke University School of Medicine (Table 1). These specimens had no prior thoracic surgery and were without other traumatic or iatrogenic disturbance beyond age-appropriate degenerative changes.
Age (years) and gender of the cadaveric specimens
| Cadaver # | Male | Cadaver # | Female |
|---|---|---|---|
| MD161129119 | 66 | MD16112392 | 73 |
| VA16120510 | 71 | MD161126104 | 71 |
| MD16120511 | 64 | MD16121652 | 53 |
| CT16092693 | 71 | MD161126106 | 65 |
| MD161128113 | 81 | MD16121441 | 71 |
| Cadaver # | Male | Cadaver # | Female |
|---|---|---|---|
| MD161129119 | 66 | MD16112392 | 73 |
| VA16120510 | 71 | MD161126104 | 71 |
| MD16120511 | 64 | MD16121652 | 53 |
| CT16092693 | 71 | MD161126106 | 65 |
| MD161128113 | 81 | MD16121441 | 71 |
Age (years) and gender of the cadaveric specimens
| Cadaver # | Male | Cadaver # | Female |
|---|---|---|---|
| MD161129119 | 66 | MD16112392 | 73 |
| VA16120510 | 71 | MD161126104 | 71 |
| MD16120511 | 64 | MD16121652 | 53 |
| CT16092693 | 71 | MD161126106 | 65 |
| MD161128113 | 81 | MD16121441 | 71 |
| Cadaver # | Male | Cadaver # | Female |
|---|---|---|---|
| MD161129119 | 66 | MD16112392 | 73 |
| VA16120510 | 71 | MD161126104 | 71 |
| MD16120511 | 64 | MD16121652 | 53 |
| CT16092693 | 71 | MD161126106 | 65 |
| MD161128113 | 81 | MD16121441 | 71 |
Bilateral gross (Zeiss Loupes: EM ProS 4.0x/450) and stereo-microscopic (Nikon SMZ 745 Stereo Microscope) dissections of the T4-T8 dorsal rami were performed by three investigators (MAH, JT, ABT). The dorsal rami were dissected with the cadavers in the prone position. The skin of the back was removed, and the trapezius and rhomboid muscles were dissected from their spinal attachments. Lateral and medial branches of the dorsal rami were traced proximally to their origins by carefully resecting the erector spinae (iliocostalis, longissimus, spinalis) and transversospinalis (semispinalis, multifidus, rotatores) musculature in the thoracic region. After the medial and lateral branches of the dorsal rami were traced to their originations, the putative articular nerves were identified and digitally photographed.
The nerve structures then were labeled with radio-opaque, 20-gauge wire. Subsequently, C-arm (AJ, MS, JT, MAH) fluoroscopic (General Electric, OEC 9900) images were obtained to delineate the fluoroscopic anatomy.
Results
A total of 100 pairs of T4-T8 spinal nerves were dissected. The posterior rami arose from the spinal nerves in their respective intervertebral foramina. Each posterior ramus was divided into medial and lateral branches near the lateral margin of the foramen. The medial and lateral branches then traveled dorsolaterally in the intertransverse space, above the superior border of the associated transverse processes, and posterior to the superior costotransverse ligament.
A total of 99 putative articular branches were identified among the branches of the 200 spinal nerves dissected (Figure 1). These nerves are referred to as putative because, although they had the characteristics of articular nerves upon dissection, histological examination was not undertaken to confirm that they were, indeed, nerves in every instance.
Origins of the T4-T8 articular branches. Following division of the thoracic spinal nerves into anterior and posterior rami, the posterior ramus of each spinal nerve passed dorsolaterally to the transverse process and divided into medial and lateral branches. The majority of articular branches derived from the posterior ramus.
Articular branches arose close to the intervertebral foramen and required partial, lateral laminectomy for full exposure in all specimens. In 87 of the 99 specimens (88%), the articular branches arose from the posterior ramus (Figure 2a and b).
a) Digital and (b) fluoroscopic images of cadaver #MD161128113. a) Digital photograph shows a close-up view of a right T4 posterior ramus (PR) and its branches, viewed from a superior-oblique angle, displaying the medial (mb) and lateral (lb) branches traced back to their origins from the PR and putative superior (to the T3–4 zygapophysial joint) and inferior (to the T4-5 zygapophysial joint) articular branches originating from the T4 posterior ramus. The black arrow points to partially removed lamina. Varying degrees of laminectomy were required to expose the articular branches in 100% of cases. b) Corresponding fluoroscopic image of right T4 posterior ramus (PR) and putative superior and inferior articular branches labeled with radiopaque wires. its = intertransverse space; tp = transverse process.
In 11 (11%) of specimens, they arose from the medial branch (Figure 3a and b).
a) Digital and (b) fluoroscopic images of cadaver #MD161128113. a) Digital photograph shows a close-up of a right T7 medial branch (mb), viewed from a superior-oblique angle, displaying the putative superior (to the T6-7 zygapophysial joint) and inferior (to the T7-8 zygapophysial joint) articular branches originating from the medial branch. The black arrow points to partially removed lamina. b) Corresponding fluoroscopic image of T7 posterior ramus (PR), its medial branch (mb), and putative superior and inferior articular branches labeled with radiopaque wires. its = intertransverse space; tp = transverse process.
The source of origin of one articular branch was considered indeterminate.
Of the 99 articular branches, 51 (52%) were superior articular branches and 48 (48%) were inferior articular branches. Superior articular branches passed to the rostral end of the zygapophysial joint behind the intervertebral foramen. Inferior branches passed to the caudal end. The articular branches were commonly suspended in the intertransverse space near the lateral aspect of the foramina, and they contacted an osseous structure, such as the pedicle, superior or inferior articulating processes, or transverse process, in only 39% of cases (Figure 4).
Fluoroscopic image of nerve structures labeled with radiopaque wires. There were no cases in which the articular branches made contact with any clinically relevant osseous target point. The articular branches are in very close proximity to the intervertebral foramina. tp = transverse process.
In six of 27 pairs (22%), there was side-to-side asymmetry in the source of the articular branch. Whereas the source of the articular branch could be the posterior ramus on one side, it was the medial branch on the other side. A “descending branch,” as found in a previous study [23], was not identified in any specimen.
Discussion
Investigations into the innervation of the midthoracic z-joints have produced discrepant findings. The seminal investigation identified the medial branch as the source of a short ascending articular branch and a descending articular branch that crosses the superolateral corner of the transverse process [22]. The latter serves as the anatomical basis for the current guideline-endorsed technique for the anesthetization of the thoracic z-joints [24]. However, a limitation to this technique was disclosed within the guideline: that the T4-T8 medial branches are frequently suspended in soft tissue, complicating the placement of a needle, which is facilitated by an osseous target point [22,24]. Subsequent work by Ishizuka et al. [23] observed that the articular branches to the thoracic z-joints could be traced, in half of cases, to a hitherto unidentified “descending branch” of the posterior ramus. None of Ishizuka et al.’s [23] specimens documented a medial branch to be the source of the articular branch to the thoracic z-joint. This further clouds the current paradigm of blocking the thoracic medial branches as having an appropriate anatomical basis for anesthetizing the thoracic z-joints.
Our data add to this body of knowledge. First, similar to Ishizuka et al. [23], we found relatively few (11%) articular branches emanating from the medial branches. Our data showed the majority of articular branches to arise directly from the posterior ramus (88%). Although we did not identify a “descending branch” in any specimen, it is possible that the articular branches identified in our specimens correspond to the “descending branch” identified by Ishizuka et al. [23]. Chua and Bogduk’s specimens showed descending articular branches off of the medial branch to cross the superolateral corner of the transverse process [22]. In contrast, our specimens showed articular branches to more commonly arise directly from the posterior ramus and much closer to the intervertebral foramina. In fact, the proximity to the foramina was such that a partial laminectomy was required to expose the articular branches in all specimens.
There appear to be several clinical consequences to the current data. First, given our observations that the articular branches most commonly arise from the posterior rami, medial branch block does not appear to have appropriate anatomical face validity to anesthetize the midthoracic z-joints. We recognize the controversial nature of this statement and note that Ishizuka et al.’s [23] findings essentially lead to the same conclusion.
Furthermore, it is reasonable to question whether a percutaneous, image-guided needle approach could block the articular branches, based upon the anatomical findings of this study. Several difficulties would likely be encountered. The articular branches were all found to be deep to the laminae, making needle access challenging. Additionally, the articular branches were within very close proximity to the intervertebral foramina. It is likely that even small volumes of local anesthetic would spread to the foraminal structures, compromising the specificity of this type of block. There is side-to-side variability in that, in 22% of pairs, the articular branches originated from the posterior ramus on one side and the medial branch on the other; given that we were only able to evaluate side-to-side asymmetry in 27 cases, we cannot rule out the possibility that this percentage represents an underestimate. Finally, and consistent with other investigations, the articular branches were most commonly suspended in the tissues of the intertransverse space. Only 39% of articular branches were observed to make contact with any clinically relevant osseous targets, such as the transverse process, the superior and inferior articular processes, or the lamina or pedicle. We believe that these points present significant obstacles to the current paradigm of anesthetization and denervation of the midthoracic z-joints.
The strengths of this study include data collected under gross, stereomicroscopic, and fluoroscopic modalities. Limitations include a small sample size of 10 cadaveric specimens and an inability to capture 100% of all articular branches, which may underestimate the individual- and population-level variation.
Conclusions
The current technique of thoracic z-joint blockade and denervation is confounded by the findings of the current study, which observes that the articular branches from T4-T8 show substantial intra- and interspecimen variability, more commonly arise from the posterior ramus rather than the medial branch, and frequently do so deep to the laminae, but without making contact with any clinically relevant osseous structures.
Acknowledgments
We wish to acknowledge the generous assistance of Mr. James Bolognesi, Manager of the Gross Anatomy Laboratory, and the late Mr. Clinton Lieweke, Manager of the Human Fresh Tissue Laboratory, both of the Duke University School of Medicine.
Additionally, we gratefully acknowledge the constructive editorial assistance of Professor Nikolai Bogduk.
References
Author notes
Funding sources: This work was funded by a research grant from the Spine Intervention Society and an internal grant from the Department of Orthopaedic Surgery at the Duke University School of Medicine.
Conflicts of interest: There are no conflicts of interest to report.
