The Spine in Three Dimensions

The spine is a stack of small bones that somehow manages to be strong enough to carry the upper body, flexible enough to let a dancer bend in half, and protective enough to shelter the spinal cord through a lifetime of movement. Nearly every topic in this book, from manipulation and posture to pain and examination, rests on knowing this structure in three dimensions. This chapter builds that mental model one layer at a time: the regions and curves, the bones, the joints and ligaments that tie them together, the muscles that move and steady them, and the neural tissue they protect.

Section 1Regions, curves and the motion segment

Five regions, thirty three bones

The vertebral column (spinal column, backbone) runs from the base of the skull to the tip of the tailbone. In a typical adult it is about 70 cm long in men and about 60 cm in women, and roughly a quarter of that length is made up of intervertebral discs rather than bone. It is divided into five regions, and the commonly taught counts are worth memorizing exactly:

• Cervical: 7 vertebrae (C1 to C7), forming the neck.
• Thoracic: 12 vertebrae (T1 to T12), each carrying a pair of ribs.
• Lumbar: 5 vertebrae (L1 to L5), the large weight bearing bones of the low back.
• Sacral: 5 vertebrae that fuse in adulthood into one sacrum.
• Coccygeal: usually 4 small segments (the range is 3 to 5) that fuse partly or completely into the coccyx.

That gives 33 vertebrae at birth and, because the sacral and coccygeal segments fuse, about 24 separate movable vertebrae above the sacrum in the adult (7 + 12 + 5). Natural variation is common. Some people have a sixth lumbar segment, some have a lowest lumbar vertebra partly fused to the sacrum (sacralization), and some have a first sacral segment that behaves like a lumbar vertebra (lumbarization). These are called transitional segments, and they matter in practice because they can change how vertebral levels are counted on an image. A practitioner who assumes every spine has exactly five lumbar vertebrae can mislabel a level.

Curves: primary and secondary

Seen from the front or back, a healthy spine is straight. Seen from the side, it forms a gentle S shape made of four sagittal curves. Two curves are concave forward, called kyphoses (singular kyphosis): the thoracic and the sacral. Two curves are concave backward, called lordoses (singular lordosis): the cervical and the lumbar. The terms describe the shape of a normal curve; in everyday clinical language, "kyphosis" or "lordosis" is also used loosely for an exaggerated curve, so read the context carefully.

The kyphotic curves are called primary curves because they are present before birth. The fetus is curled into a C shape, and the whole column is one long kyphosis. The lordotic curves are called secondary curves (or compensatory curves) because they develop after birth in response to function. The cervical lordosis begins to form in the first months as the infant lifts the head against gravity, and it becomes well established when the child can sit with the head upright. The lumbar lordosis forms later, as the child begins to crawl, stand and walk, roughly around the end of the first year, and it continues to mature through childhood. A useful way to remember the order is to follow the milestones from the top down: head control first, then standing.

Secondary curves are produced mostly by wedging of the discs rather than of the bones. The front of a cervical or lumbar disc is taller than the back. In the thoracic region the reverse is partly true of the vertebral bodies themselves, which are slightly taller behind than in front, so the thoracic kyphosis is more "built into" the bone. This is one reason the thoracic curve changes less with posture than the lumbar curve.

Why curves help

A curved column sounds weaker than a straight one, but the alternating curves serve several purposes. They allow the column to flex slightly under load and spring back, spreading force through discs, ligaments and muscles instead of concentrating it on bone. They keep the head and trunk balanced over the pelvis so that the line of gravity passes close to the joints that carry it, which reduces the muscle work needed to stand. They also enlarge the thoracic and pelvic cavities.

Normal curve size varies widely between healthy people. Commonly cited standing values are roughly 20 to 45 degrees of thoracic kyphosis and roughly 40 to 60 degrees of lumbar lordosis, but ranges differ by measurement method, age and sex. A curve outside a reference range is a finding to interpret, not a diagnosis on its own; many people with large or small curves have no pain, and many people with pain have average curves. Abnormal curvature in the frontal plane, a sideways curve with rotation, is scoliosis, discussed in later chapters on posture and examination.

The motion segment

Because the spine is made of many repeating units, it is useful to study the smallest unit that shows the column's behavior. This unit is the motion segment, also called the functional spinal unit (FSU). It consists of two adjacent vertebrae plus all the soft tissue between them: the intervertebral disc, the two facet joints, and the ligaments that connect the two bones. The muscles that cross the segment are usually discussed separately, but they act on it constantly. Physicians and researchers describe motion as occurring "at L4 to L5," meaning in the motion segment formed by L4, the L4 to L5 disc, and L5.

At each typical motion segment the two vertebrae meet at three joints. In front is the interbody joint, the vertebral bodies joined by the disc, a cartilaginous joint (specifically a symphysis). Behind are the paired zygapophyseal joints, usually called facet joints, which are synovial joints. Together they form the three joint complex. The idea behind the name is that these three joints share the load and share the motion, so a change in one affects the others. When a disc loses height, for example, the facets are pushed closer together and carry more load, which can contribute to facet joint wear over time. This is the logic behind the commonly described degenerative cascade, a sequence in which disc and facet changes influence each other over decades.

Anterior and posterior columns

A practical way to think about the motion segment is to split it into an anterior column and a posterior column. The anterior column (bodies and discs) is built for compression: it carries most of the axial load in upright standing and absorbs shock. The posterior column (arches, facets, spinous processes and their ligaments) is built for guidance and tension: the facets steer motion and resist shear and rotation, while the posterior ligaments resist being pulled apart in flexion. Surgeons and radiologists use a related three column model for spinal fractures (anterior, middle and posterior), in which the middle column is the back half of the body and disc plus the posterior longitudinal ligament. Injuries that damage two or more columns are generally considered more likely to be unstable.

Section 2Vertebrae by region

The typical vertebra

Almost every vertebra follows the same basic plan, and regional differences are variations on that plan. A typical vertebra has two main parts. In front is the vertebral body (centrum), a short cylinder of spongy (cancellous) bone wrapped in a thin shell of compact bone. The body is the weight bearing part. Behind is the vertebral arch (neural arch), a ring of bone that encloses the vertebral foramen. Stacked together, the vertebral foramina form the vertebral canal, which houses the spinal cord, its coverings and the nerve roots.

The arch has named parts that you should be able to point to:

• Pedicles: two short, thick pillars that connect the arch to the back of the body. The upper and lower edges of each pedicle have notches; the inferior notch of one vertebra and the superior notch of the one below make the roof and floor of the intervertebral foramen.
• Laminae: two flat plates that run from the pedicles backward and toward the midline, meeting to close the arch.
• Spinous process: a single process projecting backward from the junction of the laminae. It is the bump you feel down the middle of the back.
• Transverse processes: paired processes projecting laterally from the junction of pedicle and lamina.
• Articular processes: two superior and two inferior processes, each with a smooth cartilage covered facet. The inferior articular processes of one vertebra meet the superior articular processes of the one below to form the facet joints.
• Pars interarticularis: the narrow strip of lamina between the superior and inferior articular processes. A stress fracture here is spondylolysis, most common at L5.

The spinous and transverse processes act as levers for muscles and attachment points for ligaments. Longer levers give muscles more mechanical advantage.

Cervical vertebrae

The neck has the smallest and most mobile vertebrae. All seven cervical vertebrae share one feature found nowhere else: a hole in each transverse process called the transverse foramen (foramen transversarium). The vertebral artery usually enters the transverse foramen at C6 and ascends through C1 before turning medially over the posterior arch of the atlas and entering the skull through the foramen magnum. The C7 transverse foramen is small and usually carries only small veins. The typical cervical vertebrae (C3 to C6) also have:

• Small, wide, rectangular bodies with raised lateral lips on the upper surface called uncinate processes. These form the uncovertebral joints (joints of Luschka) with the body above. They guide motion and help keep the disc from bulging laterally, and bony spurs from them can narrow the intervertebral foramen in older adults.
• A large, triangular vertebral foramen to fit the cervical enlargement of the cord.
• Short, often bifid (split) spinous processes, typically C2 to C6.
• Transverse processes ending in anterior and posterior tubercles, separated by a groove for the spinal nerve. The anterior tubercle of C6 is especially large and is called the carotid tubercle.
• Articular facets set at roughly 45 degrees, facing upward and backward on the superior processes, which allows a generous range of flexion, extension, lateral flexion and rotation.

Three cervical vertebrae are atypical. The atlas (C1) has no body and no spinous process. It is a ring made of an anterior arch, a posterior arch and two lateral masses that carry kidney shaped superior articular facets for the occipital condyles. The atlanto-occipital joint allows the nodding "yes" motion, flexion and extension of the head. The axis (C2) has a tooth shaped projection, the dens (odontoid process), rising from its body. Developmentally, the dens is the body of C1 that fused to C2. The dens sits against the back of the anterior arch of the atlas and is held there by the strong transverse ligament of the atlas. The atlanto-axial joints allow the "no" motion: about half of all cervical rotation occurs here. The axis also has the first large, palpable spinous process below the skull. Finally, C7 is called the vertebra prominens because its long, non bifid spinous process is usually the most visible bump at the base of the neck (although T1 is sometimes as prominent, which is why palpation counts need confirmation).

Thoracic vertebrae

The defining feature of thoracic vertebrae is their relationship with the ribs. Each typical thoracic vertebra has costal facets (demifacets) on the side of the body where rib heads articulate, and transverse costal facets on the transverse processes where the rib tubercles articulate. A typical rib head (ribs 2 to 9) sits across the disc and touches two vertebrae: the body of its own number and the body above. Thus rib 6 articulates with T5 and T6 and with the transverse process of T6. Ribs 1, 10, 11 and 12 articulate with only one vertebral body, and ribs 11 and 12 have no joint with a transverse process.

Other thoracic features include heart shaped bodies of intermediate size, a small round vertebral foramen (the cord is narrowest here, and the canal leaves little spare room), and long spinous processes that slope steeply downward, especially in the middle thoracic region (about T5 to T8), where the tip of a spinous process can lie at the level of the body below. The thoracic facets are nearly vertical and lie close to the frontal (coronal) plane, facing backward on the superior processes. That orientation favors rotation and lateral flexion, but the rib cage stiffens the region overall.

A palpation guideline often taught for the thoracic spine, the rule of threes, estimates how far below its own transverse processes each spinous process tip lies; studies have questioned its precision, so treat it as a rough guide.

Lumbar vertebrae

The lumbar vertebrae carry the most body weight and have the largest bodies, which are wide and kidney shaped when viewed from above. They have no transverse foramina and no costal facets. The spinous processes are thick, square (hatchet shaped) and point almost straight backward. The transverse processes are long and slender. On the back of each superior articular process is a small bump called the mammillary process, and at the base of the transverse process is a smaller accessory process; both anchor the deep multifidus and intertransversarii muscles.

The lumbar facets are curved and oriented mostly in the sagittal plane in the upper lumbar spine, becoming more coronal (more turned to face front and back) at L4 to L5 and L5 to S1. The superior facets are concave and face medially and backward; the inferior facets are convex and face laterally and forward. This interlocking shape allows flexion and extension readily but limits axial rotation to only a few degrees per segment. The more coronal orientation at the bottom helps resist the forward shear that develops where the lumbar spine sits on the tilted top of the sacrum. L5 is the largest lumbar vertebra and is somewhat wedge shaped, taller in front than behind, contributing to the lumbosacral angle.

Sacrum and coccyx

The sacrum is a large triangular bone formed by fusion of five sacral vertebrae, usually completed by about the mid twenties. It sits wedged between the two ilia, forming the back of the pelvis. Key landmarks include:

• The sacral base (the top surface of S1), whose front edge projects forward as the sacral promontory.
• The ala (wings) on each side of the base.
• Four pairs of anterior sacral foramina and four pairs of posterior sacral foramina, through which the ventral and dorsal rami of S1 to S4 exit.
• The median sacral crest (fused spinous processes) on the back.
• The sacral hiatus, the lower opening of the sacral canal where the laminae of S5 (and sometimes S4) fail to meet, flanked by the sacral cornua. Clinicians use the hiatus for caudal epidural injections.
• The auricular surfaces (ear shaped) on each side, which form the sacroiliac joints with the ilia.

The sacral base is tilted forward. The angle between the top of S1 and a horizontal line, the sacral base angle (Ferguson's angle), is commonly taught as about 30 to 40 degrees in standing, although reported values vary. A steeper sacrum generally goes with a deeper lumbar lordosis. The coccyx, the tailbone, is a small triangle of usually four rudimentary segments. It anchors the gluteus maximus, parts of the levator ani and the anococcygeal ligament, and it bears some weight when a person sits leaning back. Pain here is called coccydynia.

Section 3Discs, facets and ligaments

The intervertebral disc

There are 23 intervertebral discs in the adult spine. The first lies between C2 and C3; there is no disc between the occiput and the atlas or between the atlas and the axis. The last true disc lies between L5 and S1, and a small rudimentary disc remnant may be present between the sacrum and coccyx. Disc height increases from the neck to the low back, and the discs together make up roughly a quarter of the height of the column above the sacrum. Because discs lose a little water under load during the day and regain it lying down overnight, people are measurably taller in the morning than in the evening, often by about 1 to 2 cm.

Each disc has three parts that work together as a pressurized system.

• Nucleus pulposus: the gel-like center. In a young adult it is about 70 to 90 percent water, held there by proteoglycans (especially aggrecan), large molecules whose negatively charged glycosaminoglycan chains attract water by osmosis. The nucleus also contains loose type II collagen. Because it is mostly water, it is nearly incompressible; when loaded it pushes outward in all directions, converting vertical compression into tension in the surrounding annulus. In the lumbar spine the nucleus sits slightly behind the center of the disc.
• Annulus fibrosus: the tough outer ring of about 15 to 25 concentric layers called lamellae. The outer annulus is rich in type I collagen (tensile strength); the inner annulus has more type II collagen and proteoglycan and blends into the nucleus. The collagen fibers in each lamella run at an angle of roughly 60 to 65 degrees from vertical, and the angle alternates direction in successive layers, like the plies of a radial tire. This crisscross arrangement lets the annulus resist twisting in both directions and contain the pressure of the nucleus. The posterior and posterolateral annulus is thinner than the anterior annulus, which is one reason disc herniations usually occur in a posterolateral direction.
• Vertebral endplates: thin layers of hyaline cartilage (about 1 mm thick) that cover the top and bottom of each disc and attach it to the vertebral bodies. The cartilage endplate rests on a perforated plate of bone, the bony endplate, with a raised rim called the ring apophysis. The endplate is the weakest part of the segment under pure compression, so a large compressive load tends to crack the endplate before it ruptures a healthy annulus. A herniation of nucleus through the endplate into the vertebral body is called a Schmorl's node.

Blood supply, nutrition and nerves of the disc

The adult disc is the largest structure in the body without its own blood supply. Small vessels reach only the outermost annulus. The cells of the nucleus and inner annulus are fed by diffusion: nutrients such as glucose and oxygen pass from capillaries in the vertebral bodies through the endplates, and waste such as lactic acid diffuses out. The trip can be several millimeters, which is long for diffusion. Movement and changes in load are thought to help by pumping fluid in and out, which is one reason activity is often described as good for disc health. When the endplate calcifies with age, diffusion slows, the disc cells struggle, proteoglycan content falls, and the nucleus loses water. These changes are part of normal aging and are very common on imaging in people without pain.

Innervation is also limited to the periphery. The outer third of the annulus receives sensory fibers from the sinuvertebral nerve (recurrent meningeal nerve), which re-enters the intervertebral foramen to supply the posterior annulus, posterior longitudinal ligament and dura, and from branches of the gray rami communicantes on the front and sides. In degenerated discs, nerve fibers can grow deeper into fissures of the annulus, a finding that has been linked with discogenic pain.

The facet joints

The zygapophyseal (facet) joints are true synovial joints, with articular cartilage, a synovial membrane, synovial fluid and a fibrous joint capsule. Many contain small inward folds called meniscoids (synovial folds or fat pads) that fill the gaps at the edges of the joint. The capsule is richly supplied with sensory receptors, including mechanoreceptors and nociceptors, so the facet can be a pain source and also a source of position information. Each facet joint is innervated by the medial branch of the dorsal ramus from the level at and the level above the joint. For example, the L4 to L5 facet joint receives medial branches from the L3 and L4 dorsal rami. This dual supply is why diagnostic medial branch blocks target two nerves for one joint.

Facet orientation is the single biggest reason each region moves differently. In the cervical spine the facets are angled about halfway between horizontal and vertical, so the neck can move in all directions. In the thoracic spine the facets lie in the frontal plane, which favors rotation and lateral flexion. In the lumbar spine they lie mostly in the sagittal plane and curve around each other, which favors flexion and extension and blocks rotation. Chapter 2 develops these motion patterns in detail.

The facets share load with the disc. In upright standing with a normal lordosis they are generally said to carry a modest fraction of the axial compressive load (figures from about 10 to 20 percent are commonly quoted), and this share rises in extension and when the disc loses height. Their larger roles are to resist shear, limit rotation and guide the path of motion.

Ligaments of the spine

Ligaments are bands of dense connective tissue connecting bone to bone. In the spine they limit motion at the end of range, protect the cord, and, because they contain sensory receptors, contribute to the body's sense of position. Two run along the bodies, and several connect the arches.

• The anterior longitudinal ligament (ALL) is a broad, strong band on the front of the bodies and discs, from the occiput and anterior tubercle of the atlas to the sacrum. It limits extension and is thick enough to resist forward bulging of the disc.
• The posterior longitudinal ligament (PLL) runs on the back of the bodies, inside the vertebral canal, from C2 (where it continues upward as the tectorial membrane) to the sacrum. It limits flexion. It is narrower than the ALL, and in the lumbar region it narrows further and is attached mainly to the discs in a pattern that leaves the posterolateral corners less covered. This is another reason herniations tend to go posterolateral rather than straight back.
• The ligamentum flavum (yellow ligament) connects the laminae of adjacent vertebrae. It is unusual for its high elastin content, which gives its yellow color and lets it stay taut without buckling into the canal when the spine extends. It limits flexion. With aging it may thicken and fold, narrowing the canal and contributing to spinal stenosis.
• The interspinous ligaments join adjacent spinous processes and limit flexion.
• The supraspinous ligament joins the tips of the spinous processes from C7 down to the sacrum. Above C7 it expands into the ligamentum nuchae, a midline sheet that anchors neck muscles.
• The intertransverse ligaments join transverse processes and limit lateral flexion to the opposite side.
• The facet capsules are ligaments too, and they limit flexion and rotation.

The upper cervical spine has special ligaments that protect the brainstem. The transverse ligament of the atlas holds the dens against the anterior arch; with vertical bands above and below it forms the cruciform ligament. The paired alar ligaments run from the sides of the dens to the occipital condyles and limit rotation of the head on the neck. The tectorial membrane covers these ligaments from behind. Rheumatoid arthritis and Down syndrome can weaken the transverse ligament, allowing atlantoaxial instability, a contraindication to forceful upper neck manipulation.

Section 4Muscles and fascia of the trunk

Three layers of back muscles

Back muscles are easiest to learn in layers, sorted by depth, function and nerve supply. The superficial and intermediate layers are called extrinsic back muscles: they lie on the back but developed with the limbs and ribs, and they are supplied by ventral rami (or, for the trapezius, the spinal accessory nerve). The deep layer contains the intrinsic or true back muscles, which move and stabilize the spine itself and are supplied by dorsal rami of the spinal nerves. That one distinction, ventral versus dorsal rami, explains a great deal of clinical anatomy.

• Superficial layer (moves the upper limb): trapezius, latissimus dorsi, levator scapulae, and rhomboid major and minor.
• Intermediate layer (thin respiratory muscles): serratus posterior superior and serratus posterior inferior, which attach to ribs.
• Deep layer (moves and stabilizes the spine), itself arranged from superficial to deep as the splenius group, the erector spinae, the transversospinalis group, and the small segmental muscles.

The intrinsic muscles

The splenius capitis and splenius cervicis wrap like a bandage (splenius means bandage) from the spinous processes of the lower neck and upper thorax upward and laterally to the skull and upper cervical transverse processes. Acting on one side they rotate and side bend the head toward the same side; acting together they extend it.

The erector spinae is the large muscle mass that forms the two ridges beside the midline of the low back. It arises from a broad common tendon attached to the sacrum, iliac crest and lumbar spinous processes, and it splits into three columns. From lateral to medial they are the iliocostalis, longissimus and spinalis, remembered by the phrase "I Love Spine." Working together on both sides, the erector spinae extend the spine and, importantly, control flexion by contracting eccentrically as we bend forward. Working on one side, they side bend the trunk toward the same side.

The transversospinalis group lies deep in the groove between the spinous and transverse processes. Its fibers run upward and medially from transverse processes to spinous processes. Three muscles make up the group, getting shorter and deeper as you go: semispinalis (spans about four to six segments, most developed in the neck and upper thorax), multifidus (spans about two to four segments, thickest in the lumbar region) and rotatores (span one or two segments, best developed in the thoracic region). On one side they rotate the trunk toward the opposite side. The lumbar multifidus deserves special attention because it has a large cross section close to the joints it crosses, which suits it to controlling small segmental motions. Imaging studies have reported reduced size and increased fat in the multifidus of many people with chronic low back pain, although whether this is a cause or a result of pain is still studied.

The deepest segmental muscles are the interspinales and intertransversarii. They are tiny and contain a very high density of muscle spindles, so they are thought to act mainly as position sensors rather than movers. In the upper neck, four pairs of suboccipital muscles (rectus capitis posterior major and minor, obliquus capitis superior and inferior) connect the occiput, atlas and axis. They fine tune head position, are supplied by the suboccipital nerve (dorsal ramus of C1), and bound the suboccipital triangle, through which the vertebral artery runs.

The abdominal wall and other trunk muscles

The front and sides of the trunk are closed by four abdominal muscles. The rectus abdominis runs vertically from the pubis to the lower ribs and xiphoid, divided by tendinous intersections, and is the main trunk flexor. The external oblique runs downward and forward ("hands in your pockets"), the internal oblique lies beneath at roughly right angles to it, and the transversus abdominis, the deepest, runs horizontally around the waist like a belt. Rotation pairs the external oblique of one side with the internal oblique of the other: turning the trunk to the right uses the left external oblique and right internal oblique. The transversus abdominis increases intra-abdominal pressure and tensions the thoracolumbar fascia. All four are supplied by the ventral rami of the lower thoracic nerves (about T7 to T12), with L1 contributing to the lower internal oblique and transversus.

Several other muscles complete the trunk "cylinder." The diaphragm forms its roof, and its crura attach to the upper lumbar bodies. The pelvic floor forms its base. The quadratus lumborum forms part of the back wall of the abdomen, and the psoas major runs along the sides of the lumbar bodies to the femur. In the neck, the longus colli and longus capitis (the deep neck flexors) lie on the front of the cervical bodies, while the scalenes and sternocleidomastoid move the neck and assist breathing.

The thoracolumbar fascia

The thoracolumbar fascia (lumbodorsal fascia) is a tough sheet of connective tissue that wraps the deep back muscles. In the lumbar region it has three layers. The posterior layer covers the back of the erector spinae and attaches to the spinous processes. The middle layer attaches to the tips of the lumbar transverse processes and separates the erector spinae from the quadratus lumborum. The anterior layer covers the front of the quadratus lumborum. The layers fuse at the lateral edge of the erector spinae to form the lateral raphe, which is where the transversus abdominis and internal oblique attach. Superficially, the latissimus dorsi takes origin from the posterior layer, and below, the fascia blends with the gluteus maximus fascia across the midline.

This arrangement makes the thoracolumbar fascia a link between the arms, trunk and legs. When the latissimus dorsi on one side and the gluteus maximus on the other contract together, as in walking or throwing, they tension the fascia diagonally across the low back. When the transversus abdominis contracts, it pulls on the fascia laterally and stiffens the lumbar segments.

Section 5The spinal canal, cord and nerve roots

The spinal cord and its coverings

The spinal cord begins at the foramen magnum as a continuation of the medulla and, in most adults, ends at about the level of the L1 to L2 disc as a cone-shaped tip, the conus medullaris. At birth it reaches lower, about L3, because during development the column grows faster than the cord. The cord is about 45 cm long in the adult and is not uniform: it widens at the cervical enlargement (cord segments about C4 to T1, supplying the upper limbs) and the lumbosacral enlargement (about L1 to S3 segments, supplying the lower limbs), where more neurons serve the limbs.

Three membranes, the meninges, surround the cord. The dura mater is the tough outer tube. Unlike the cranial dura, the spinal dura is separated from bone by the epidural space, which contains fat and a valveless internal vertebral venous plexus (part of the network often called Batson's plexus). Inside the dura is the delicate arachnoid mater, and beneath it the subarachnoid space filled with cerebrospinal fluid (CSF). The pia mater clings to the cord surface. Pia forms the denticulate ligaments, tooth-like lateral extensions that anchor the cord to the dura, and continues below the conus as the filum terminale, which anchors the cord to the coccyx. The dural sac itself ends at about S2.

Below the conus, the vertebral canal contains a bundle of lumbar, sacral and coccygeal nerve roots descending to their exits, called the cauda equina ("horse's tail"). Because there is no cord below about L2, a needle can be placed in the subarachnoid space at L3 to L4 or L4 to L5 to sample CSF (lumbar puncture) with little risk to the cord. A large central disc herniation or other mass in the lower canal can compress many roots at once, producing cauda equina syndrome: saddle area numbness, new bladder or bowel dysfunction, and leg weakness. It is a surgical emergency and one of the most important red flags in spine care.

Spinal nerves and how they are numbered

There are 31 pairs of spinal nerves: 8 cervical, 12 thoracic, 5 lumbar, 5 sacral and 1 coccygeal. Each forms from a dorsal (sensory) root, which carries the cell bodies of sensory neurons in the dorsal root ganglion, and a ventral (motor) root. The roots join in or just beyond the intervertebral foramen to form a short mixed spinal nerve, which divides almost at once into a dorsal ramus (to the deep back muscles, facet joints and skin of the back) and a larger ventral ramus (to the limbs and the front and sides of the trunk, forming the cervical, brachial and lumbosacral plexuses). The sinuvertebral nerve branches back into the canal.

Numbering trips up many students because the neck has 8 nerves but 7 vertebrae. The rule is:

• In the cervical spine, each nerve exits above the vertebra of the same number. C1 exits between the occiput and atlas, C6 exits between C5 and C6, and so on.
• C8 has no vertebra of its own and exits between C7 and T1.
• From T1 downward, each nerve exits below the vertebra of the same number: the L4 nerve exits through the L4 to L5 foramen.

Because the cord ends high, there is also a growing mismatch between cord segment level and vertebral level. In the lower thoracic region, the cord segments sit about two to three levels above the vertebrae of the same number, and all of the lumbar and sacral cord segments are packed into the region behind roughly T11 to L1. The lumbar and sacral roots must therefore descend a long way through the canal before they exit.

The intervertebral foramen and lateral recess

The intervertebral foramen (neural foramen) is the doorway through which each spinal nerve leaves the canal. Its boundaries are:

• Roof and floor: the pedicles of the vertebrae above and below.
• Front: the back of the vertebral bodies and the disc (and, in the neck, the uncovertebral joint).
• Back: the facet joint and its capsule, with the lateral edge of the ligamentum flavum.

Each of these boundaries can narrow it. A posterolateral disc bulge or herniation, an osteophyte from the uncovertebral or facet joint, a thickened ligamentum flavum, or loss of disc height that lowers the pedicle above can all reduce the space, a condition called foraminal stenosis. In the lumbar spine the nerve root occupies only the upper part of the foramen, beneath the pedicle, so loss of disc height pinches it from below and in front.

Just before a lumbar root enters its foramen, it travels in a gutter along the side of the canal called the lateral recess (subarticular zone), bounded in front by the body and disc, laterally by the pedicle, and behind by the superior articular process. Enlargement of that facet can trap the traversing root here. This geography explains a key clinical rule. At the L4 to L5 disc, the L4 root is already exiting high in its foramen, while the L5 root is traversing the canal on its way down to the L5 to S1 foramen. A typical posterolateral herniation at L4 to L5 therefore compresses the L5 root, the traversing one. A far lateral (foraminal) herniation at the same disc compresses the exiting L4 root. In the cervical spine the roots exit almost horizontally, so a C5 to C6 herniation usually affects the C6 root.

Blood supply of the cord

The cord is supplied by one anterior spinal artery, which runs in the front midline and feeds roughly the front two thirds of the cord, and two posterior spinal arteries, which feed the back third, including the dorsal columns. These longitudinal channels are reinforced along the way by segmental (radicular) arteries entering through the foramina. The largest is the great anterior radicular artery, the artery of Adamkiewicz, which usually enters on the left between about T9 and T12. The mid thoracic cord is a watershed zone with relatively little collateral flow. Veins drain into the valveless vertebral venous plexuses, which connect with pelvic and abdominal veins; this connection is one proposed route for the spread of prostate and other cancers to the vertebrae.

Words worth knowing