Sœmmerring’s error: the root of the story. The C8 nerve is nonsense. A historical review and anatomical perspectives

Patrick Chaynes ^1,2* Luana Carfagna ^1,3 Marina Poinsignon ² Amaury De Barros ^1,2,4*

• ¹ Laboratoire d'Anatomie, Faculté de Santé, Université Toulouse III Paul Sabatier, Toulouse, France
• ² Service Neurochirurgie, Pôle Neurosciences CHU Toulouse, Toulouse, France
• ³ Service de Chirurgie Pédiatrique, Centre Hospitalier Universitaire de Toulouse, Toulouse, Occitanie, France
• ⁴ INSERM U1214 Centre d'Imagerie Neuro Toulouse (ToNIC), Toulouse, France

Which doctor did not learn the arrangement of the spinal nerves? (I did, sure you did too). Which medical student did not ask: "How can there be eight cervical nerves when there are only seven cervical vertebrae?" (I did, maybe you too). The answer is always the same: "My dear young student, this discrepancy is easily explained; the first cervical nerve runs superior to the first cervical vertebra, atlas, and the last runs between the seventh cervical and the first thoracic vertebrae, leaving six cervical nerves in between: so there are eight in number."In the same way, the 42nd British edition of Gray's Anatomy (Standring, 2020), as well as every single anatomical textbook worldwide, reports 31 pairs of spinal nerves arranged in: 8 cervical, 12 thoracic, 5 lumbar, 5 sacral and 1 coccygeal nerves. This fundamental assumption is so deeply ingrained that it would be a sacrilege to have doubt about as it appears to be incontestable certainty. Anyone who dares to contradict it should fear being either presumptuous or foolhardy, but that ordering, established in the late 18th century by Samuel T. Soemmerring (1798), requires it.As we now consider 31 pairs of spinal nerves, this was not always the case. In the mid16th century, Vesalius described 30 pairs of spinal nerves (Vesalius, 1555). These corresponded to the 7 cervical, 12 thoracic, 5 lumbar and 6 sacral vertebrae, and the suboccipital nerve passing between the occipital bone and the first cervical vertebra was joined to the last pair of cranial nerves. This description was so authoritative that it was found in Willis (1664) and Vieussens (1684) until the late 18th century (Monro, 1782;Winslow, 1732).At the end of the 18th century and the beginning of the 19th, the suboccipital nerve has in turns been considered as a cranial nerve (Bichat 1812; Monro, 1782; Sabatier 1791) or a spinal nerves (Asch, 1750;Bell, 1812;Haller, 1778;Portal, 1803) depending on whether the author mainly considered its cranial aspect (arising from a single root of the medulla oblongata) or its spinal one (arising from the spinal cord along two roots). Since all anatomists asserted the first thoracic nerve exits below the first thoracic vertebra because it gives rise to the first intercostal nerve which runs under the first rib, when the suboccipital nerve was considered as a spinal nerve (Asch, 1750;Bell, 1812;Haller, 1778), the cervical nerves which already were seven in number raised to eight and Soemmerring was the first to describe the 8 th cervical spinal nerve.Soemmerring named the spinal nerves according to their point of issue from the vertebral column (Soemmerring, 1798). Their segmental pattern is transferred by the somite (Keynes & Stern, 1984) which also gives rise to the vertebra, and that of neural arches develops in interaction with the spinal ganglia and nerves (Christ & Wilting, 1999). All spring from the spinal cord by a dorsal and a ventral root that join and pass the vertebral column through a foramen limited by pedicles of two adjacent vertebrae; some of them (i.e. the first seven) lay superior to the vertebra whereas others (i.e. thoracic and lumbar nerves) pass from below (Standring, 2020).One question arises: if spinal nerves reflect segmental organisation of the nervous system and human frame, how can somites and nerves, two structures so dependent on each other, have different relationships over the vertebral column? As the shape and composition of all spinal nerves are identical and mirror the anatomical organisation of the body, the answer becomes obvious: to one somite, one nerve, one vertebra:Where is Soemmerring's error?Soemmering seems to be wrong because he described 8 cervical spinal nerves rather than 7 due to suboccipital nerve described as a spinal and not a cranial nerve. The problem is that he did not question the origin of the first thoracic nerve which, for ancient anatomists, must necessarily have been the origin of the first intercostal nerve.All spinal nerves pass through an intervertebral foramen and divide in two rami in very close contact to the vertebral arch of the underlying vertebra (Lazorthes, 1971). The first spinal nerve courses over the posterior arch of the first cervical vertebra at the level of the groove of the vertebral artery, which is homologous to the superior notch of the vertebral pedicle (Testut, 1893), especially since sometimes a thin bony spiculum placed above converts it into a complete foramen (Standring, 2020;Testut, 1893). The second spinal nerve runs through a fairly wide space serving as an intervertebral foramen. It consists of the very discreet inferior notch of the atlas and the pedicle of the axis which merge with the lamina (Testut, 1893). The next 23 spinal nerves exit the vertebral column between the superior and inferior vertebral notches and divide in two rami. The anterior ramus of the third to seventh spinal nerve lies on the sulcus nervi spinalis on the upper surface of the underlying transverse process, and the dorsal ramus crosses its posterior root and wraps around the cranial articular process of the same vertebra; those of the eighth are closely applied on the neck of the first rib (Paturet, 1964) before the posterior ramus wraps the cranial articular process of the first thoracic vertebra (Lazorthes, 1971). In the thoracic region, the anterior rami give rise to the intercostal nerves which pass on the neck of the underlying rib whereas the dorsal rami run posteriorly and enter a vertical osteofibrous slot limited by the overlying transverse process superiorly, the upper edge of the underlying rib inferiorly and the underlying cranial articular process medially. In a same way, the dorsal rami of the lumbar nerves run posteriorly downwards over the transverse process of the underlying vertebra and wrap around its superior articular process. The caudal spinal nerves run through the sacral canal. On each side, four T shape conduits separated by a sagittal bony column homologous to pedicles allow the 26th to 29th spinal nerves to pass through an intervertebral foramen.The last two pairs of spinal nerves exit the sacral canal through the caudal opening (Poirier & Charpy, 1899). The penultimate, more lateral, passes in front of the horn of the sacrum between the lateral edge of the sacrococcygeal joint and the medial bundle of the lateral sacrococcygeal ligament which often ossifies with age. This osteofibrous passage limited by the 5th sacral vertebra and the first coccygeal piece serves as the penultimate intervertebral foramen (Testut, 1893). The last spinal nerve, very delicate, exits medially from the previous one to obliquely run along the dorsal surface of the first coccygeal piece before passing under the dorsal ligament of the mid-coccygeal joint (Testut, 1893). The last intervertebral foramen is therefore limited by the first two pieces of the coccyx converted into an osteofibrous foramen by this ligament.Since every spinal nerve course through a foramen, and its posterior ramus wraps the cranial articular process of the underlying vertebra, it is understood there is a close relationship between one segment of the peripheral nervous system and the underlying segment of the vertebral column. All spinal nerves course superior to their corresponding vertebra.Such as all vertebrates, human body exhibits a segmental organization. This is mostly visible in the metameric division of the nervous system and the series of bones that forms the vertebral column. The number of spinal nerves is strictly correlate to the number of vertebrae. The presence of supernumerary vertebrae leads to an increase in number of spinal nerves and conversely, their reduction in number will cause a decrease in that of spinal nerves (Paturet, 1964).Segmental organization is a fundamental process remarkably conserved during phylogenesis since it also exists in invertebrates (earthworms). In human embryogenesis, somites occur very early in development (Christ & Wilting, 1999). The segmental pattern is transferred by the somite to the blood vessels, the spinal ganglia, and the nerves 11 according a hierarchy in which the somite represents the primary, the spinal ganglia and nerves the secondary, and the vertebral neural arches the tertiary elements (Christ & Wilting, 1999). The metameric division of the neural tube, the vertebral column, the thorax, and therefore, the ribs depends on their arrangement.Each pair of somites is centred by a pair of spinal nerves and connected to a segment of spinal cord. The cells of the early somite are assigned to different compartments: the dorsal half gives rise to the dermomyotome out of which subcutaneous tissue and skeletal musculature proceed (Christ & Wilting, 1999) while the ventral half is source of the sclerotome which forms an important component of the axial skeleton (Figure 1a). In turn, the sclerotome can be subdivided into a ventral, dorsal, lateral and central compartments that will form the vertebral body, the main part of the neural arch, the distal part of the rib, the pedicle including the transverse process and the proximal part of the rib, respectively (Christ & Wilting, 1999;Scaal, 2016). It is furthermore divided by a transient faint cleft that not extend to the axial area (Ebner, 1888) into a dense caudal and a loose cranial half. The spinal nerves and ganglia develop into the loose cranial half of the sclerotome (Keynes & Stern 1984; Remak, 1855; Scaal, 2016) (Figure 1b).While two pairs of adjacent sclerotomes contribute to the formation of the vertebral body by fusion of their inferior part to the superior part of those of next caudal sclerotome by the "so-called" phenomenon of resegmentation (Ebner, 1888), formation of the lateral vertebral elements (especially transverse process and proximal part of the rib) originates from the sole dense caudal half of the sclerotome 18 and takes place without resegmentation (Christ & Wilting, 1999) (Figure 1b).Since each developing spinal nerve develops in the loose cranial half, it lies on the dense caudal half of its sclerotome which gives rise to the ventral part of the pedicle, transverse process and proximal part of the rib (the costal head and neck, including the costovertebral joint) (Scaal, 2016). To resume, a sclerotome issue from a somite is finally divided by an intervertebral disc and correspond to the distal part of a vertebra (the loose cranial half part of the sclerotome) and the cranial part of the following vertebra with its transverse process or proximal rib (the dense caudal part of the sclerotome). Hence, whatever the level of the vertebral column, the spinal nerve corresponding vertebra is the one of the transverse process or neck of the rib of which it courses on (Figure 1c).The location of the boundary between head and neck seems to vary according to authors (Christ & Wilting, 1999;Padget, 1954 Scaal, 2016). Its cranial half gives rise to the proatlas, which will later be incorporated into the apical part of the dens axis, when the caudal half fuses with the cranial part of the underlying somite to form the dens axis (the vertebral body of atlas) which later fuses with the vertebral body of the second vertebra, the axis (Christ & Wilting, 1999). Since the first spinal nerve centres the first spinal somite it is located on the dense zone of the first sclerotome, the second spinal nerve runs on the dense caudal zone of the second spinal sclerotome at the origin of the neural arch of axis. Once it is understood that each spinal nerve runs over its corresponding vertebra, for example the eighth spinal nerve with the eighth vertebra, which is nothing else than the first thoracic vertebra: the current name, e.g. "eighth cervical spinal nerve" becomes inadmissible.Since there are 31 pairs of spinal nerves, if the eighth pair of spinal nerves is the first thoracic one, there must be two pairs of coccygeal nerves.The neural tube extends longitudinally along the axis of the embryo. Even though there can be up to a total of 38 or 39 pairs of somites (O'Rahilly et al., 1990), several disappear in relation to caudal regression of the human embryo. In some instance rudiments of coccygeal nerves still remain detectable (O'Rahilly & Müller, 2003). The caudalmost part of the primitive spinal cord, pia mater, arachnoid mater and dura mater corresponds to the very end of the vertebral column, and in most cases attaches to the fifth coccygeal vertebra (O'Rahilly et al., 1990): the spinal nerves emerge perpendicularly through their own intervertebral foramen and there is no cauda equina. As the vertebral column elongates, the spinal cord ascends, the caudal spinal nerves become stretched vertically until they reach their foramen to exit: the cauda equina appears. Arising at the tip of the conus medullaris, the filum terminale attaches to the posterior surface of the first two coccygeal pieces (Christ & Wilting, 1999). This caudal insertion marks the position of the lower end of the definitive spinal cord and the last spinal nerve (the 31st) face the second coccygeal vertebra. Effectively, in the adult, the last two pairs of spinal nerves exit the vertebral column through the sacral hiatus.The penultimate, more lateral, crosses the sacral horns and passes through an osteofibrous slot. The latter serves as a vertebral pedicle, being limited by the fifth sacral vertebra and the first coccygeal piece, especially when a ligament that often ossifies with age converts it into a complete penultimate foramen. The last pair runs along the dorsal surface of the first coccygeal piece before exiting laterally under the dorsal ligament of the mid-coccygeal joint (Keynes & Stern, 1984). This last intervertebral foramen converted into an osteofibrous foramen by this ligament is limited by the first piece of the coccyx superiorly and the second inferiorly. Obviously, the last two spinal nerves exit the vertebral column each above a different coccygeal piece. The first coccygeal nerve courses on the first coccygeal vertebra and the second nerve on the second piece courses through a foramen.According to these anatomical and embryological arguments: the thoracic nerve run over its corresponding rib, the eighth spinal nerve corresponds to the first thoracic nerve and there are two coccygeal nerves. The eighth cervical nerve does not exist, this implies the metamere of the spinal cord and the 31 pairs of spinal nerves innervating striated muscles (myotomes) and skin (dermatomes), must be arranged in: 7 cervical, 12 thoracic, 5 lumbar, 5 sacral and 2 coccygeal. c. Projection of the sclerotome boundaries on the differentiated vertebral column. The spinal nerve (Yellow) courses on the transverse process or proximal rib (Dark pink) of its corresponding vertebra (Grey). Dark grey corresponding to caudal dense half of the sclerotome and light grey corresponding to loose cranial half of the sclerotome. Due to resegmentation, Sclerotome is not corresponding to vertebra but is centred of a vertebral disc (Middle blue)