Introduction
Thoracic pain manifesting as mid-back or upper-back pain is the least common of all disability-related chronic spinal problems with a prevalence of 13% compared to 32% for neck pain and 43% for low back pain. Prevalence of thoracic facet pain is 42% in patient population with localized thoracic paraspinal pain. Thoracic facet syndrome was first described in 1987. Dreyfuss and Fukui described the pain referral pattern from thoracic facet joints from various levels. The facet joints are synovial joints and are affected by osteoarthritis like any other synovial joints in the body, along with other inflammatory disorders. Mechanical injury resulting in capsular tears, capsular avulsions, subchondral fractures, intraarticular hemorrhage, and fractures of the intraarticular process all can lead to chronic pain. Clinical features are not very helpful in diagnosing facet mediated pain. The symptomatology is not much different from those suffering from myofascial pain or disc pathology. Radiological tests are also not helpful in clarifying the diagnosis, and there is no correlation between the extent of changes seen on the images and the severity of clinical symptoms. The controlled nerve block is the only reliable way to diagnose thoracic mediated facet pain with 80% pain relief from the block as a diagnostic criterion. There is Level II evidence for the controlled block in diagnosing thoracic facet joint as a pain generator while the evidence is only Level IV for long-term analgesic outcome after thoracic radiofrequency neurotomy.
Anatomy
Thoracic facet joints are well innervated. The anatomy of the medial branch of the posterior ramus and the innervation of the facet joints was explored in detail by Chua and Bogduk in their well- and extensively referenced 1995 paper. Using an X40 dissecting microscope, a total of 84 medial branches in four embalmed human adult cadavers were studied. Ishizuka et al. explored the course of 120 pairs of thoracic nerve and its branches in 10 cadavers adding to the knowledge gleaned by Chua et al. A thoracic spinal nerve divides into anterior and posterior rami. The anterior ramus continues as intercostal nerve and is around 2.5 mm to 3 mm in diameter. The posterior ramus runs laterally and gives rise to three branches: the descending branch, the medial branch, and the lateral branch. The descending branch is the first branch of the posterior ramus and descends medially. Two types of descending branches have been described: the articular type when the branch can be seen directly innervating the facet joint capsule and the nonarticular type when the nerve supplies the muscle fibers around the facet joint but innervates the joint capsule through the fine muscular fibers. The medial branch of the ramus comes off within 5 mm from the lateral margin of the neural foramen. The medial branch then courses laterally, away from the neural foramen, in an inferodorsal trajectory posterior to superior costotransverse ligament. Once it reaches the superolateral corner of the transverse process, it turns medially, wrapping around the lateral border of the multifidus muscle. It then runs along the posterior surface of the tip of the transverse process, in the plane between the origin of the multifidus muscle medially and the origin of the semispinalis thoracis muscle laterally from the transverse process. The nerve runs between the two muscles dorsally and medially innervating the two muscles. At the upper thoracic level, the medial branch ends up supplying the overlying skin. This prototype anatomical course of the medial branch is present at T1–T4 and T9–T10 level. At other thoracic levels the course varies from this prototypical course. The medial branch at T11 hugs the base of the superior articular facet of T12, while the medial branch of the T12 posterior ramus runs at the junction of the superior articular facet and L1 transverse process ( Fig. 6.1 ).
At the mid-thoracic levels (T5–T8), the course of the medial branch is quite different and more variable, as the branch never touches the transverse process at these levels. The nerve after coming off the posterior ramus runs laterally in the intertransverse space, then hooks along the multifidus muscle lateral border to turn medially and runs toward the facet joint, lying on the multifidus, never touching the transverse process. Essentially, the branches are suspended in the space between the intertransverse spaces.
At all the levels the medial branch innervates the facet joint by terminating as an ascending articular branch and a descending articular branch. The descending branch comes off the nerve as it crosses the superolateral corner of the transverse process and supplies the joint at the level, while the ascending branch supplies the joint above. Thus, it is essential to perform medial branch blocks at the superolateral corner of the transverse process, before the nerve divides, to capture the descending branch ( Fig. 6.2 ).
Radiological anatomy
The radiological appearance of the thoracic spine is complicated by the presence of ribs. The pedicle, transverse processes, and facet joints are not easy to see under flouro. The important differences are:
- 1.
The thoracic spinous processes are often hard to see and often deviate off midline and are not suitable to determine the degree of obliquity. Therefore, they are not as reliable a landmark here as in the lumbar area.
- 2.
The pedicles are directed upward and are not seen very clearly in an anteroposterior (AP) view as the trajectory does not align with the x-ray beam in this view. The rounded shadow seen on an AP flouro view of the thoracic spine close to the upper endplate of the vertebral body is casted by costovertebral joints and the pedicles.
- 3.
The thoracic transverse process is a short but a robust structure that has broad attachment to the inferior facet pillar. The orientation of the processes is directed backward and upward. This oblique angulation requires contralateral oblique orientation of the C-arm to see the ipsilateral transverse process clearly.
- 4.
The facet joints are oriented in the frontal plane, and the articular pillars are wider compared to the lumbar level, extending laterally beyond the vertebral bodies.
Diagnosis
Thoracic facet disease is not a clinical diagnosis, nor can it be confirmed using CT or MRI. Common features of localized pain and tenderness are vague, and there are no physical tests that can be performed to evaluate the facet joints in the thoracic spine. The diagnosis can only be confirmed by blocking the medial nerve branch of the dorsal ramus. Each joint is innervated by the medial branch from the same level along with the nerve from one level above, thus each joint requires two blocks at two levels. A single diagnostic injection has high false-positive rate (34% to 48%). Currently, facet joint pain is diagnosed when a patient meets the following criteria:
- 1.
Eighty percent or more pain relief from a set of controlled comparative local anesthetic blocks using low volume (0.5 ml). Usually, 1% lidocaine and 0.25% bupivacaine are used for comparative blocks.
- 2.
Pain relief should last longer with bupivacaine than with lidocaine. It should be 2 h with lidocaine and at least 3 h with bupivacaine.
- 3.
Maneuvers or activities that were painful before should become very tolerant for the duration of the local anesthetic after the block.
- 4.
The two blocks are done 3 to 4 weeks apart.
The validity of the comparative block in minimizing false positive outcome is also dependent on prevalence of the disease. With a prevalence rate of 46% in patients presenting to the pain clinic with localized thoracic pain, positive response to two comparative local anesthetic blocks means that physicians can be 80% confident that the patient has the facet disease. But this also means that there is a 20% chance that it is not a facet disorder despite having a good response to the two facet diagnostic blocks, and ablative therapy of the nerves performed on the basis of good response will not provide long-term relief in these properly selected patients.
Diagnostic block steps
The medial nerve branch block is usually done using fluoroscopic guidance. Image guidance is a must for a safe performance of this procedure.
Consent: After informed consent, patient is taken to the procedural suite. The risks are quite minimal and are needle-related and/or medication-related. Needle-related risks include pain from the procedure, infection, bleeding, dural puncture, pneumothorax, and nerve damage. Medication injection rarely causes an issue as very low volume is used but can be an issue if injected intrathecally by accident. If particulate steroid is being used, it can lead to hypopigmentation if injected too close to the skin accidently and rarely spinal cord infarction if embolism happens.
Intravenous Access: This is not needed if sedation is not planned.
Position: The patient is placed in a prone position.
Target Levels: A thoracic spine pain pattern is used to pick potential painful joints. To block each joint, two medial nerve branches are blocked. For T1/T2 joints, the block is performed at T1 and T2 transverse processes and so on ( Table 6.1 ).
