Innervation and Pain Patterns of the Thoracic Spine Nikolai Bogduk

Innervation and CHAPTER Pain Patterns of the Thoracic Spine Nikolai Bogduk During the life of the first two editions of this book, few, if any, definitive studies were published on the nature, origin, diagnosis, or treatment of mechanical, or idiopathic, thoracic spinal pain. This problem remains underserved by the literature. Such infor- mation that might be harvested on this topic still sterns from seminal studies under- taken more than 40 years ago. There have, however, been certain developments in the modern era. At last, formal studies of the innervation of the thoracic spine have been conducted'r': and modern studies of pain patterns have been undertaken using radio- logically controlled techniques.' INNERVATION The thoracic spine is innervated in a manner similar to that of the cervical and lumbar spines. The posterior elements (those structures that lie behind the intervertebral fo- ramina) are innervated by the thoracic dorsal rami. The anterior elements (which lie anterior to the intervertebral foramina and spinal nerves) are innervated by the sinu- vertebral nerves. THORACIC DoRSAL RAMI Each thoracic dorsal ramus arises from its spinal nerve and passes directly posteriorly, entering the back through an osseoligamentous tunnel bounded by a transverse pro- cess, the neck of the rib below, the medial border of the superior costotransverse liga- ment, and the lateral border of a zygapophyseal joint (Figure 5-1). The nerve then runs laterally through the space between the anterior lamella of the superior costo- transverse ligament anteriorly, and the costolamellar ligament and the posterior la- mella of the superior costotransverse ligament posteriorly (Figure 5-1). It divides in this space, some 5 mm from the lateral margin of the intervertebral foramen, into a medial and a lateral branch.' From its origin the medial branch passes slightly dorsally and inferiorly, but largely laterally, within the intertransverse space. There, it is embedded in areolar tis- 73

74 Chapter 5 Innervation and Pain Patterns of the Thoracic Spine .\".~~---8 v.:~-10 Figure 5-1 The innervation of the thoracic spine as viewed from the rear. On the left, the vertebral laminae have been resected to reveal the contents of the vertebral canal. The dural sac has been retracted to demonstrate the thoracic sinuvertebral nerves. On the right, the courses of the thoracic dorsal rami are shown. For clarity, muscles such as the levatores costarum and iliocostalis have not been depicted. 1, Semispinalis thoracis; 2, multifidus; 3, lateral costo- transverse ligament covering costotransverse joint; 4, posterior lamella of the superior costo- transverse ligament; 5, costolamellar ligament; 6, anterior lamella of the superior costotransverse ligament; 7, nerve to costotransverse joint; 8, medial branch of dorsal ramus; 9, articular branches to zygapophyseal joints; 10, lateral branch of dorsal ramus; 11 and 12, medial and lateral slips of longissimus thoracis; 13, branches of sinuvertebral nerves to epi- dural vessels; 14, sinuvertebral nerve; 15, spinal nerve; 16, branches of sinuvertebral nerve to dura mater; 17, sinuvertebral nerve; 18, branches to posterior longitudinal ligament; 19, ra- dicular artery. sue and accompanied by small arteries and veins. Opposite the tip of the transverse process, the medial branch curves dorsally around the lateral border of the posterior lamella of the superior costotransverse ligament and aims inferiorly for the superolat- eral corner of the transverse process. It enters the posterior compartment of the back by crossing this corner and running caudally along the posterior surface of the tip of the transverse process, through the cleavage plane, between the origin of multifidus medially and that of the semispinalis laterally.I Covered by the semispinalis, the me-

Innervation 75 dial branch curves inferiorly and medially over the dorsal aspect of the fascicles of multifidus, to which it supplies multiple filaments. Other branches supply the semi- spinalis. At upper thoracic levels, a long branch continues over the dorsal surface of the multifidus toward the midline, where it penetrates the fascicles of spinalis thora- cis, splenius cervicis, rhomboids, and trapezius to become cutaneous.\" At lower thoracic levels, the medial branches of the dorsal rami retain an exclusively muscular distribution. Each medial branch furnishes ascending and descending articular branches to the zygapophyseal joints. 1 Ascending branches arise from the medial branch as it passes caudal to the zygapophyseal joint above. These branches are short and ramify in the inferior aspect of the joint capsule. Slender, descending branches arise from the me- dial branch as it crosses the superolateral comer of the transverse process. They fol- Iowa sinuous course between the fascicles of multifidus to reach the superior aspect of the capsule of the zygapophyseal joint below.1 The capsules of the joints are en- dowed with free nerve endings and mechanoreceptors, although the latter are more sparse than in the cervical and lumbar zygapophyseal joints.i The lateral branches of the thoracic dorsal rami continue the projection of the dorsal ramus in the intertransverse space, initially running parallel to the medial branches before they enter the posterior compartment of the back. Beyond the tip of the transverse process, each lateral branch descends caudally and laterally, weaving between the fascicles of the longissimus thoracic muscle (Figure 5-1). As a rule, each nerve supplies the fascicles of longissimus that attach to the transverse process and rib above the level of origin of the nerve and sometimes the fascicle from the rib next above.' Continuing caudally and laterally, the lateral branches enter and supply the il- iocostalis muscles. The lateral branches of the lower thoracic (T7 to T12) dorsal rami eventually emerge from the iliocostalis lumborum to become cutaneous.\" Those from higher levels have an entirely muscular distribution. Articular branches to the costo- transverse joints arise from the lateral branch just above each joint where the medial branch leaves the lateral branch? (Figure 5-1). THORACIC SINUVERTEBRAL NERVES The thoracic sinuvertebral nerves are recurrent branches of the thoracic spinal nerves. Each nerve arises from two roots-a somatic root and an autonomic root. The so- matic root arises from the anterior surface or superior border of the spinal nerve just outside the intervertebral foramen. It passes into the intervertebral foramen, running in front of or sometimes above the spinal nerve, and joins with the autonomic root af- ter a course of about 2 to 3 mm. 5,7 The autonomic root arises from the grey ramus communicans at each segmental level or, in some cases, from the sympathetic gan- glion nearest the spinal nerve. 5,7 Having been formed, each sinuvertebral nerve passes through the intervertebral foramen and enters the vertebral canal, embedded amongst the branches of the segmental spinal artery and the tributaries of the spinal vein, an- terior to the spinal nerve. In the intervertebral foramen the nerve gives rise to fila- ments that supply the vertebral lamina and a branch that crosses the upper border of the neck of the nearby rib to supply the periosteum of the neck. 5,7 Other branches are distributed to the vessels within the vertebral canal. Terminal branches ramify in the anterior surface of the vertebral laminae, the dural sac, and the posterior longitudinal ligament.l'\" As in the cervical and lumbar spine, the thoracic spine is innervated by dense mi- croscopic plexuses that accompany the posterior and anterior longitudinal ligaments\" (Figure 5-2). The posterior plexus is derived from the thoracic sinuvertebral nerves;

76 Chapter 5 Innervation and Pain Patterns of the Thoracic Spine all A . .. '. .0.: :• • • ' ••• \" ••• f • -v , •• • .'.',',''..... , B Figure 5-2 The pattern of innervation of the thoracic vertebral bodies and intervertebral discs, as seen in human fetuses. A, Transverse section. Branches to the anterior longitudinal ligament (all) emanate from the sympathetic trunks (st) and sympathetic ganglia (sg). Branches to the pos- terolateral and posterior aspects of the intervertebral disc (ivd) stem from the sympathetic trunk and from the sinuvertebral nerves (svn) that are directed to the posterior longitudinal ligament. drg, Dorsal root ganglion; dr, dorsal ramus; vr, ventral ramus; drr, dorsal root; vrr, ventral root. B, Longitudinal view showing the posterior longitudinal plexus. The sinuverte- bral nerves (svn) form a dense plexus that ramifies over the back of the intervertebral discs (ivd). p, Location of pedicles; drg, dorsal root ganglion. (Based on Groen GJ, Baljet B, Drukker J: AmJAnal 188:282, 1990.1

Patterns of Pain 77 the anterior plexus from the thoracic sympathetic trunks and rami communicantes. Each plexus furnishes branches that supply the longitudinal ligaments and branches that penetrate the vertebral bodies and intervertebral discs. Branches from the poste- rior plexus innervate the ventral aspect of the dural sac.\" SOURCES OF PAIN The structures that receive an innervation, and therefore are possible sources of pain in the thoracic region, are the thoracic vertebrae, the dura mater, the intervertebral discs and longitudinal ligaments, the posterior thoracic muscles, the costotransverse joints, and the thoracic zygapophyseal joints. Of these structures, studies have dem- onstrated the ability of the thoracic muscles, the thoracic discs, and the zygapophyseal joints to produce thoracic spinal pain. When injected with hypertonic saline in normal volunteers, the thoracic interspi- nous muscles and ligaments produce local and referred pain across the posterior chest wall.9,l o When provoked by discography, the thoracic discs produce posterior thoracic pain in normal volunteers and reproduce pain in patients with posterior thoracic pain. II In normal volunteers, distending the thoracic zygapophyseal joints with injec- tions of contrast medium produces posterior thoracic pain.' Thoracic spinal pain can be relieved by intraarticular injections ofbupivacaine and triamcinolone into the zyga- pophyseal joints, 12 PATIERNS OF PAIN In the thoracic region the phenomenon of referred pain poses more diagnostic diffi- culties than in any other region of the vertebral column. Of foremost importance is the diagnosis of visceral pain referred to the chest wall. Chest pain can be caused by cardiac, pulmonary, and pleural disease, as well as diseases of the mediastinum, esophagus, and diaphragm. Because many of these diseases are potentially life threat- ening, they must be recognized and managed or specifically excluded. Specialist con- sultation may be required for this purpose. However, preoccupation with visceral dis- ease has led to the neglect and even denial of the possibility that chest wall pain may be somatic or skeletal in origin, Experiments in normal volunteers have shown that noxious stimulation of the in- terspinous structures at thoracic levels can produce somatic referred pain to both the posterior and anterior chest wall.9•IO This pain follows somewhat of a segmental pat- tern (Figure 5-3), insomuch as stimulation of higher levels causes referred pain at higher levels in the chest wall. However, there is no consistent location for referred pain from a particular segment. The locations ascribed to different segments differ in different studies, and the segmental pattern is not always strictly sequential. Stimula- tion of a particular segment may cause referred pain at a higher level than stimulation of the segment below (Figure 5-3). Because of these variations and irregularities, the location of any referred pain cannot be used to deduce the exact segmental location of its source. Similar patterns of pain arise from the thoracic zygapophyseal joints' (Figure 5-4). The pattern is again quasisegmental; the higher the source, the higher the loca- tion of the referred pain; but again, the patterns overlap. In principle, however, it would appear that the segmental location of a patch of posterior thoracic spinal pain might be identified from the distribution of pain with an accuracy of perhaps plus or minus one or two segments.

78 Chapter 5 Innervation and Pain Patterns of the Thoracic Spine Figure 5-3 Referred pain patterns in the chest. The shaded areas illustrate the distribution of referred pain reported by normal volunteers after stimulation of interspinous structures at the seg- mental levels indicated. The figures on the left are based on the data of Kellgren.\" Those on the right are based on Feinstein et al.lO Note the differences in the distribution of pain in the two sets of figures, and the extensive overlap in distribution shown in the figures on the right.

Pathology 79 T4-5 T3-4 T6-? T5-6 T8-9 T7-8 T10-11 T9-10 Figure 5-4 Referred pain patterns of the thoracic zygapophyseal joints. IFrom Dreyfuss P, Tibiletti C, Dreyer Sj: Spine 19:807, 1994.1 PATHOLOGY There is an absence of any pathological data, or even circumstantial evidence, to ex- plain why thoracic discs, muscles, zygapophyseal joints, or costovertebral joints might be a source of thoracic pain. It is therefore difficult to state with any certainty what disorders might afflict these various structures to produce pain. Infective and neoplastic diseases of the bone can affect the thoracic vertebrae, but such conditions are usually evident on radiological investigations and are not likely causes of idiopathic thoracic pain. Neoplastic disorders of the thoracic dura or epi- dural blood vessels typically manifest themselves by causing symptoms of spinal cord compression and have not been described as causing pain without neurological signs. Herniation of thoracic intervertebral discs is an uncommon disorder but is usually at- tended by signs of nerve root irritation or spinal cord compression.l ' Furthermore, thoracic disc herniations most commonly occur at lower thoracic levels (T9 to TIO) and are associated with pain in the lumbar region and abdominal wall rather than in the chestY Very few pathological conditions have been described as affecting the posterior thoracic muscles and the synovial joints of the thoracic spine. The costotransverse and thoracic zygapophyseal joint can be involved in ankylosing spondylitis, but it would be unusual for this condition to be the source of thoracic pain in the absence of signs of concomitant involvement of the sacroiliac joints or other features of ankylosing spon-

80 Chapter 5 Innervation and Pain Patterns of the Thoracic Spine dylitis. Rheumatoid arthritis can affect the costotransverse joints14•15 and may spread from these sites to involve the adjacent intervertebral discs.\" Rheumatoid arthritis is also recognized as affecting the thoracic zygapophyseal joints, although not as severely as the costotransverse joints.IS The German literature'\" describes degenerative joint disease of the thoracic zygapophyseal and costotransverse joints, but there is no physi- ological evidence of whether or not all joints so affected become painful. A notion attractive to manual therapists is that either thoracic zygapophyseal joints or the costotransverse joints can be affected by mechanical disorders that cause pain and are amenable to manipulative therapy. However, the pathology of these pu- tative disorders is not known. Further study of these conditions depends critically on the development and implementation of diagnostic blocks of these joints. The issue of muscular pain is even more speculative. The thoracic spine is abun- dantly covered by posterior muscles, and the thoracic transverse processes and ribs are virtually riddled with muscle attachments. Moreover, much of the posterior thoracic musculature is formed by lumbar muscles inserting into thoracic levels or cervical muscles arising from thoracic levels. This arrangement invites the suggestion that thoracic pain arising from muscles could be caused by cervical or lumbar disorders that disturb the normal function of these overlapping muscles. Spasm of these muscles as a result of cervical or lumbar pain or excessive tension in them as a result of abnor- mal lumbar or cervical mechanics could be perceived as straining their thoracic at- tachments and thereby causing pain. However, clinical or physiological evidence sub- stantiating any of these concepts is lacking. No studies have shown that anesthetizing certain muscle insertions relieves thoracic pain secondary to cervical or lumbar dis- eases, nor have any controlled studies verified that correction of abnormal cervical or lumbar postures relieves thoracic muscular pain. These various reservations and seemingly negative conclusions should not be in- terpreted as denials of the possibility that idiopathic thoracic pain can be caused by disorders of the thoracic synovial joints or muscles. Indeed, the analogy with the lumbar and cervical regions makes it likely that at least the zygapophyseal and costo- transverse joints would be potent sources of otherwise undiagnosed thoracic pain. However, what is emphasized is the absence, to date, of any definitive clinical, experi- mental, or pathological data that permits endorsement of this notion. DISCUSSION In a sense this chapter may not seem helpful for readers hoping to find explanations and answers to thoracic pain problems, because the conclusions made are so diluted with reservations. However, this accurately reflects the state of the art with respect to idiopathic thoracic pain. In the absence of appropriate anatomical, experimental, and clinical data, one cannot make legitimate conclusions. There is a dire need for basic data in this field. References 1. Chua WH, Bogduk N: The surgical anatomy of thoracic facet denervation, Acto Neurochir 136:140,1995. 2. McLain RF, Pickar ]G: Mechanoreceptor endings in human thoracic and lumbar facet joints, Spine 23:168, 1998. 3. Dreyfuss P, Tibiletti C, Dreyer S]: Thoracic zygapophyseal joint pain patterns: a study in normal volunteers, Spine 19:807, 1994.

References 81 4. Johnston HM: The cutaneous branches of the posterior primary divisions of the spinal nerves, and their distribution in the skin,] Anat Physio/43:80, 1908. 5. Hovelacque A: Anatomic des nerfs Craniens et Rachidiens et du System Grand Sympathique, Paris, 1927, Doin. 6. Wyke BD: Morphological and functional features of the innervation of the costovertebral joints, Folia Morphol Praba 23:286, 1975. 7. Hovelacque A: Le nerf sinu-vertebral, Ann Anat Path Medico-Chir 2:435, 1925. 8. Groen GJ, Baljet B, Drukker J: Nerves and nerve plexuses of the human vertebral column, Am] Anat 188:282, 1990. 9. KellgrenJH: On the distribution of pain arising from deep somatic structures with charts of segmental pain areas, Clin Sci 4:35, 1939. 10. Feinstein B, Langton JBK, Jameson RM, Schiller F: Experiments on referred pain from deep somatic tissues,] Bone Joint Surg 36A:981, 1954. 11. Wood KB, Schellhas KP, Garvey TA, Aeppli D: Thoracic discography in healthy individu- als: a controlled prospective study of magnetic resonance imaging and discography in asymptomatic and symptomatic individuals, Spine 24:1548, 1999. 12. Wilson PR: Thoracic facet joint syndrome-clinical entity? Pain Supp 4:S87, 1987. 13. Taylor TKF: Thoracic disc lesions, ] Bone Joint Surg 46B:788, 1964. 14. Weinberg H, Nathan H, Magora F et al: Arthritis of the first costovertebral joint as a cause of thoracic outlet syndrome, Clin Orthop 86:159, 1972. 15. BywatersEGL: Rheumatoid discitis in the thoracic region due to spread from costoverte- bral joints, Ann Rheum Dis 33:408, 1974. 16. Hohmann P: Degenerative changes in the costotransverse joints, Zeitshr fur Orthop 105:217,1968.