A thorough knowledge of fairly basic anatomy and neurology is necessary to enable nerve root or spinal cord disorders to be understood. Spinal disorders, including tumours, may present with combination of the following four groups of features.

Disorders of the spinal column often produce localized pain. If the disorder actually arises from the spinal column, such as a metastatic tumour in a vertebral body, localized pain may well predate other clinical features such as nerve root or spinal cord disorders. A metastatic tumour in the spine should be suspected in any patient with a known primary tumour who develops backache, often between the shoulder blades. Such patients often present late with profound and irreversible spinal cord damage. Tenderness on palpation or percussion of the spinous processes is often a very helpful physical sign. Localized deformity or absence of a spinous process on palpation are other useful signs.

A characteristic of root pain is that it is often worse by night than by day and may wake the sufferer. It is relieved by walking or sitting upright in a chair. Nerve root pain is often thought of as a shooting pain radiating in a dermatomal distribution, often associated with pins and needles, and both being made worse by coughing, sneezing, or straining. This may be so, for instance, with a lumbar disc protrusion compressing the L5 or S1 root, but root pain due to a tumour more commonly produces an aching pain which is worse on inactivity.

The features of a lower neurone lesion are wasting, weakness, hypotonia, and associated reduction of the tendon reflexes. It is important to be able to test the relevant myotomes (Table 1) 599. These have withstood the test of time and busy clinical practice, although other clinicians lists may differ. Thus a C5 root lesion will cause weakness of shoulder abduction and a C6 lesion weakness of elbow flexion and an absent or impaired biceps jerk. A C7 root lesion causes weakness of elbow extension and reduction or absence of the triceps jerk, while weakness of finger flexion denotes a C8 root lesion. A T1 lesion produces weakness of the small hand muscles. A series of tests can therefore be used to test from C5 to T1 quickly and accurately.

Testing the myotomes in the legs is a little more complicated, but hip flexion is dictated by the L2/3 myotomes, knee extension by the L3/4 myotomes, ankle inversion by L4/5, and plantar flexion by S1/2. All of the other commonly tested manoeuvres (hip extension (gluteus maximus), knee flexion (hamstrings), ankle eversion, and dorsiflexion of the foot and toes) are controlled by the L5/S1 myotomes. An L5 root lesion can be differentiated by a lesion in S1 by examining the ankle reflex; this is equivalent to plantar flexion and contains no L5 myotome. Thus absence of the ankle jerk indicates an S1 lesion; weakness of the L5/S1 myotomes and an intact ankle jerk is due to an L5 root lesion.

The maximum circumference of the calves should be measured. Any difference of more than 1 cm is suggestive of a lower motor neurone lesion, providing that there are no other obvious abnormalities.

A long established lower motor neurone lesion will produce fasciculation of the muscles of that myotome. Pain may also arise as a result of a motor root being affected, the pain following the myotome distribution of the root. This is an intense, aching type of pain, more ill-defined than sensory nerve root pain. Thus a C7 motor root pain is in the region of the pectoralis major and radiates down the back of the arm to the triceps muscle. On the left side such a root pain is not infrequently misdiagnosed as pain from myocardial ischaemia.

Sensory root symptoms, predominantly those of pain in a dermatomal distribution are often associated with pins and needles. Patients may complain of hyperalgaesia, an intense, unpleasant, and unusual pain in response to the slightest painful stimulation of the skin. Hyperesthaesia, heightened and unpleasant appreciation of light touch, may also occur. Hyperesthaesia is well known in association with herpes zoster or post-herpetic neuralgia; affected patients are usually hypoalgaesic in the hyperesthaetic area. The distribution of dermatomes is best tested where there is least overlap from adjacent dermatomes. Since three adjacent nerve roots need to be cut before dermatomal sensory loss is appreciated, because of the degree of overlap, it is more important to know where to test for each dermatome than to know the rather variable geographic boundaries of each dermatome. Table 2 600 provides a list of the more important dermatomes.

Spinal cord compression is associated with disorders of motor function, causing an upper motor neurone lesion, impaired sensation, and abnormal bladder and bowel function below the level of the lesion.

Pain rarely results from pressure on the spinal cord. Any such pain that occurs is described more as an altered and unpleasant sensation (dysaesthesia). Numbness or tingling of the legs is a common and early symptom. Patients frequently describe a feeling of a tight band around the trunk or limb; this feeling often indicates a lesion of the posterior columns of the spinal cord. Impairment of pain and temperature sensation may produce symptoms of burning or coldness, but there are often no symptoms other than complaints of painless burns or other injuries. Careful enquiry about sensory symptoms will allow the clinicians to focus sensory testing on the appropriate area.

The first motor symptom to appear is often that of ‘stiffness’ of the legs: scuffing may cause the patient to trip on uneven surfaces. Weakness and disability follows and when a marked degree of spasticity exists the patient may develop involuntary jerking of the legs. Spontaneous ankle clonus may also develop: this is characteristically aborted by plantar flexion of the foot.

Symptoms of bladder and bowel dysfunction should be sought. The most reliable symptom is lack of a normal feeling of a full bladder; the patient also fails to feel urine passing down the urethra, and may know when the bladder is empty only by cessation of the noise of urine flowing into the receptacle. Patients may also be unable to decide whether they are about to pass flatus or faeces, and males will develop impotence. These symptoms also occur in patients with cauda equina (nerve root) disorders.

Examination of a patient with spinal cord damage discloses an upper motor neurone lesion of the legs, trunk, and arms depending on the level of the lesion. The deltoid jerk, which is not a tendon reflex but a muscle reflex, should be tested. Tapping the normal deltoid muscle does not elicit a reflex contraction; in the presence of an upper motor neurone lesion a reflex response occurs, indicating that the lesion is at C4 or above.

Examination of sensation in these circumstances should be directed at finding a sensory level. This level may be abrupt and obvious when profound lesions exist, when the diagnosis has been made too late. In early disease considerable persistence may be required before a sensory level can be determined. Normally there is heightened appreciation of pinprick on passing from the abdomen to the costal margin. Patients often find it easier to appreciate a change in sensation when a pin is dragged lightly up the body. In the thoracic region a sweat level is often present, due to the body being sympathectomized below the lesion: this may be useful in an unconsciousness patient with suspected spinal cord damage. When looking for suspected dermatomal sensory loss it is better to dissuade patients from sensory signs, for if too assiduously looked for, patients will describe a normal variation of skin sensitivity. However, when looking for a sensory level one has to be most obsessive. If a possible level is found, testing further up the trunk will show whether the level is genuine or just a variation in degree of sensory loss beneath the real sensory level. The finding of a root lesion above the level denotes the true pathological level: the sensory level does not attain the pathological level until the cord compression is profound.

In early stages of cord compression the sensory loss is in the legs; this then gradually ascends up the trunk. Initially there is sparing in the buttock region, unless the tumour is at the level of a conus medullaris or cauda equina.

Intramedullary tumours or syringomyelia produce a different picture to that of other spinal tumours. These tumours first affect the spinothalamic fibres crossing in the anterior commissure, causing a ‘suspended dissociated sensory loss’. There is loss of pain and temperature sensation but not light touch at the level of the lesion. Progression of the intramedullary lesion causes caudal extension of the sensory loss, eventually producing a typical sensory level. As an intramedullary lesion is often elongated, lower motor neurone signs, especially absent tendon reflexes, are common and obvious at the level of the lesion.

It is unusual to find a pure Brown–Sequard syndrome, but the incomplete form is usually caused by a soft cervical disc protrusion, although neurologists more often see this syndrome in patients with multiple sclerosis.

It is important to examine the rest of the body. Tumours elsewhere may metastasize to the spinal column, and may also extend intradurally. Careful examination of the skin may show evidence of neurofibromatosis, while a port-wine stain, especially in a dermatomal distribution, should suggest spinal angioma. Auscultation over a spinal angioma may reveal a bruit.

The dissociation of the vertebral and the neurological levels which occurs due to the fact that the spinal cord ends at about the lower border of the L1 vertebral body should be remembered. The ‘thoracic cord’ (the T12 segment of the spinal cord) ends at the lower border of the ninth thoracic vertebral body. The lumbar plexus therefore arises in the spinal cord opposite the T10 and T11 vertebral bodies, while the sacral plexus arises opposite the T12 and L1 vertebral bodies. Failure to appreciate this dissociation will prevent investigations being performed at the appropriate spinal level (usually not high enough), and the diagnosis will not be reached. It is also worth knowing that the C8 neurological (or myelomere) level is opposite the C6/7 disc space.

Careful clinical assessment is essential, not only to determine the most relevant investigation but to ensure appropriate correlation with any abnormality seen on investigation. This warning is especially relevant in evaluation of disorders of the back or neck where ‘bulging discs’ are normal and osteophytes the rule in patients over the age of 40 years.

It is important to obtain plain radiographs at the appropriate level, after careful neurological assessment, bearing in mind the dissociation between vertebral and neurological levels. Collapse, erosion, and scalloping of the vertebral bodies (especially the pedicles) may be seen. Expansion of the intervertebral foramen, or the diameters of the spinal canal indicates a slow growing expansive lesion. Lateral radiographs of the cervical spine with the neck fully extended and flexed are particularly helpful when assessing patients with possible cervical spondylotic radiculopathy or myelopathy.

Myelography is the standard method of investigation, using a water-soluble dye injected into the subarachnoid space by lumbar or cisternal puncture. Myelography demonstrates spinal cord compression; very occasionally rapid deterioration of function may follow myelography and require immediate neurosurgical intervention. Failure to obtain CSF should suggest the possibility of a cauda equina tumour, especially if the patient experiences severe root pain during the attempted lumbar puncture. The most common error during myelography is failure to run the dye high enough. Myelography performed to investigate back pain and sciatica should routinely include examination of the midthoracic region to exclude a spinal angioma or a high intradural tumour. Use of the term ‘radiculogram’, implying mere examination of the nerve roots, is inadvisable; such an examination should never be performed. Despite the advent of MRI myelography is still the best investigation for cervical spondylotic radiculopathy.

CSF obtained during the lumbar puncture should be sent to the laboratory for analysis. If it is yellow (xanthochromic), either the patient has either suffered a subarachnoid haemorrhage, or the protein level is abnormally high, indicating a complete or near complete obstruction to CSF flow in the spinal canal.

CT scanning is of particular value for demonstrating abnormalities in the vertebrae, such as the extent of metastatic tumour and associated paraspinal spread. CT guided biopsy of neoplastic or inflammatory vertebral disease is extremely useful and avoids many fruitless operations.

CT scanning alone is of more limited value for demonstrating pathology within the spinal canal, although its value is much enhanced by administration of intrathecal contract medium. Post-myelography CT scanning is the best method of investigation for some spinal canal pathology and disc protrusions (especially thoracic disc protrusions). CT scanning alone may show a pathological lumbar disc prolapse but false-positive and false-negative results are common. MRI is much more accurate and should be used in preference to CT scanning for lumbar disc disease.

MRI is the investigation of choice for lumbar disc disease and intradural extramedullary and intramedullary lesions of the spinal cord, especially when used with gadolinium enhancement. It shows normal, degenerative, and protruding discs particularly well, and is also useful in demonstrating Arnold-Chiari malformations at the spinal–occipital junction. Calcification is not shown by MRI, and this limits its use in the assessment of bony (vertebral) disease (Fig. 4) 2310.

Spinal angiography is used to delineate spinal angiomas that have been previously demonstrated by myelography (Fig. 5) 2311. Before intravascular embolization of spinal angiomas is undertaken the exact vascular anatomy of the feeding arteries, the fistula, and the draining veins must be determined. Angiography is the only investigation which is able to do this. It is rarely used in the investigation of other spinal conditions.

The most common cause is a metastatic deposit from the lung, breast, or prostate. Primary tumours of bone, typically myeloma, or of the reticuloendothelial system such as lymphoma, are not rare (Figs. 6 and 3) 2312,2309. Less common primary tumours such as chordoma, aneurysmal bone cyst, and osteoblastomas (in children) also occur. Inflammatory diseases such as tuberculosis or staphylococcal osteomyelitis with extradural pus are still common in certain parts of the world. A thoracic disc protrusion may present as progressive spinal cord compression.

These lie within the subarachnoid space. Meningiomas, which are particularly common in the thoracic region of middle-aged women, and neurofibromas are the most common tumours in this group. A neurofibroma arising from any nerve root may extend through to the extrathecal space and on through the intervertebral foramen (dumb-bell tumour). Arachnoid cysts and dermoid cysts also occur. Arteriovenous malformations are partly extramedullary and partly intramedullary, and may also cause progressive cord damage. These may also cause sudden cord damage by haemorrhage.

These rare and slow growing tumours are usually astrocytomas and ependymomas. Both types of tumour may be cystic and require differentiation from syringomyelia.

The essential feature that should alert the clinician to spinal cord compression are symptoms or signs indicating a sensory level, especially with a progressive history. The clinical features in any one patient are a combination of vertebral or local, nerve root, and spinal cord.

Which of these groups of symptoms predominates depends on the cause of spinal cord compression. Extradural (vertebral) causes present with back pain. Intradural, extramedullary tumours present often with nerve root pain before spinal cord compression, while vertebral pain is uncommon or late. Intramedullary lesions present with a dissociated suspended sensory loss, like syringomyelia, but further progression then causes typical spinal cord compression features.

Cervical spondylotic myelopathy and acute cervical disc protrusions may mimic spinal cord compression. These will be discussed later. Spinal angiomas are extremely rare, but their diagnosis is important, since they are treatable. Acute transverse myelitis due to demyelinating disease may occasionally cause spinal cord compression with a block on the myelogram due to the acute necrosis and swelling of the spinal cord.

Plain radiographs (including the chest) often show evidence of vertebral body collapse or pedicular erosion if a metastatic lesion is present. Expansion of the spinal canal or intervertebral foramen indicates a benign tumour. Myelography is necessary for definitive diagnosis, and CT scanning delineates the extent of vertebral body or paraspinal tumour extension. MRI is the best method to delineate intramedullary (cord) lesions.

Results of treatment depend on the severity of compression; early diagnosis is essential for good recovery. The stability of the spinal column must be preserved: further spinal cord damage can occur if the spine is rendered unstable.

It is convenient to consider the stability of the spine as depending on two columns. The anterior column consists of vertebral bodies, discs, and ligaments; the posterior column consists of the facet joints, neural arch, and ligaments. Chronic instability may result from damage to just one column. Laminectomy in children, especially when performed before the growth spurt, may produce severe deformity and subsequent cord compression, and all children treated in this way should be carefully monitored. Early bracing will prevent severe deformity. ‘Acute’ instability occurs if both columns are disrupted, for instance following decompressive laminectomy in a patient with a metastatic tumour causing vertebral body collapse.

The management of spinal cord compression relies on maintenance of spinal stability. Thus extradural tumours causing vertebral body collapse should not be treated by laminectomy. A CT-guided biopsy will establish the diagnosis and radiotherapy will control the pain. If a vertebral body is affected by a slow growing and reasonably confined tumour, and the patient is otherwise fit and with some residual spinal cord function, an anterior approach to the vertebral body may be performed. Excision of the tumour and replacement of the vertebral body with either bone, acrylic, or a metal prosthesis should be combined with a stabilization procedure. Such major surgery allows reasonable recovery of neurological function.

An extradural tumour confined to the spine and lamina and producing posterior compression is properly treated by laminectomy and excision. This will not cause acute instability.

Intradural, extramedullary tumours require treatment by microneurosurgical techniques, (including bipolar coagulation). Even tumours anterior to the cord can be removed without cord damage by the use of proper techniques. Remarkable recovery can follow removal of such tumours, and surgery should always be offered, even to patients with a long history of severe spinal cord compression.

Intramedullary tumours are rare; microsurgical excision of ependymomas is being undertaken increasingly often with minimal cord damage, but radiotherapy for such tumours is also particularly effective (Fig. 9) 2315. Astrocytomas, often cystic, are less easily removed in adults but may be removed in children. These tumours grow slowly, and the choice of treatment must depend on factors such as the site and extent of the tumour, the degree of cord function, and, not least, the patient's wishes.

Neurofibromas and ependymomas (arising from the conus medullaris or the filum terminale) are the usual tumours (Figs. 7 and 8) 2313,2314. Metastatic tumours are common and generally excluded from this group, as local pain and plain radiographs of the lumbar spine quickly establish the diagnosis.

Cauda equina tumours present in different ways to different specialists. The characteristic pain is sciatica (or root pain), which wakes the patient at night, forcing him to pace about or sit in a chair. The presence of progressive nocturnal pain should suggest a tumour rather than the usual disc protrusion. Painless weakness and wasting of the legs may occur, especially if the tumour is anterior to the conus. Sensory ataxia is caused by severe loss of sense of joint position. If sphincter function is disturbed the bladder is usually affected before the bowels. Subarachnoid haemorrhage causing sudden unusual back pain, then headache and vomiting is due to the propensity of ependymomas to bleed. The cause of papilloedema is debatable; communicating hydrocephalus following subarachnoid haemorrhage is a possibility, but is more likely to be due to a disturbance of retinal drainage associated with high CSF protein levels.

The main difficulty lies in making the clinical diagnosis. Myelography confirms the diagnosis and microneurosurgical excision produces excellent results with relief of excruciating pain.

‘Osteoarthritis’ denotes degenerative arthropathy of a synovial joint, but as the only synovial joints of the spine are the apophyseal or facet joints, the term spondylosis is used generally to denote the degenerative changes seen between the vertebral bodies. Such degeneration is caused by the decrease in water content of the discs, which is seen after the age of about 20 years. The nucleus pulposus no longer acts as a ‘ball bearing’ allowing vertebrae to rotate on each other (Fig. 10) 2316: the disc space narrows and sliding motion between the vertebrae occurs. Osteophytes or bony spurs develop in response to this abnormal movement, seemingly to reduce this abnormal movement; if this movement is prevented then the osteophytes resorb. The presence of osteophytes, in conjunction with the normal movement of the spine, may produce intermittent trauma to either the nerve root (spondylotic radiculopathy) or the spinal cord (spondylotic myelopathy). A disc protrusion occurs when the nucleus pulposus herniates through a tear in the annulus; the spinal level and the position of the extrusion determines whether spinal cord or nerve root compression, or both, occurs.

Spondylosis and disc prolapse commonly occur at the most mobile section of the spine (cervical) or the point of maximum weight bearing (low lumbar region). Cervical spondylosis is more common than is cervical disc prolapse, whereas lumbar disc prolapse more commonly produces symptoms of nerve root compression than of lumbar spondylosis.

The cervical spine is a remarkably mobile structure with a normal range of movement between flexion and extension of 100 to 110 degrees. The posterior contour of the cervical spinal canal changes by 5 cm between these extremes and the cervical dura, cord, and nerve roots need to adapt to these changes. This is mainly achieved by folding or stretching like an accordion, but also by moving up and down from and into the thoracic spinal canal (Fig. 11) 2317.

Extremes of cervical flexion cause the conus to move cranially, which exacerbates sciatica, and the brain-stem to move slightly caudally, while the cervical dura becomes taut. Limitation of neck flexion therefore occurs in patients with meningeal irritation. Extension produces a slack, thicker spinal cord and nerve roots, a narrower spinal canal, due to bulging forwards of the ligamenta flava, and narrower intervertebral foramina. The degree of narrowing depends on the individual size of the spinal canal, the size of encroaching osteophytes, and the degree of extension and range of movement of the cervical spine. Thus the pathogenesis of cervical spondylosis lies in a complex of factors, but it is basically due to the presence of osteophytes and spinal movement.

The typical patient is over 50 years of age, and often nearer 60 or 70. The history is usually one of 1 to 2 years of progressive lack of use of the hands, with either numbness or tingling extending up the forearm in a glove distribution. The legs feel stiff and often scuff the ground and there is often a sense of constriction around the knee or trunk. Neck pain is rare. The symptoms may worsen steadily or intermittently, or may become static as spondylosis gradually immobilizes the neck.

The signs are those of a spastic tetraparesis with sensory signs in the legs and superimposed lower motor neurone signs in the arms, reflecting diffuse cervical cord damage. Careful neurological testing, especially of the arm reflexes, will allow correlation with the radiographic investigations. The bladder becomes affected late in the course of the disease.

Causes of spinal cord compression, especially in the region of the foramen magnum, need to be considered. Instability of the cervical spine and subsequent cord damage also occurs in patients with rheumatoid arthritis, ankylosing spondylitis, or Down's syndrome. Motor neurone disease is usually obvious.

This typically presents as a stiff neck, either on waking or after a minor twist of the neck. Pain radiates down the arm, especially during extension of the neck, producing the characteristic symptoms and signs of a root lesion.

The C6 or C7 roots are commonly affected, since these pass through the most mobile joints of the spinal column (C5/6 and C6/7). Weakness, tingling, and numbness occurs in the appropriate root distribution, and testing the arm reflexes usually allows accurate localization. Ninety per cent of patients recover within 5 or 6 weeks, but episodes commonly recur.

Other causes of nerve root compression need to be considered, especially if the symptoms do not subside. A foramen magnum tumour may inexplicably cause wasting of the small hand muscles or tingling fingers. A cervical rib (fibrous band) causes a T1 syndrome with numbness in the forearm, as does an apical lung cancer (Pancoast tumour). Ulnar neuropathy is usually characteristic. Carpal tunnel syndrome causes tingling in the thumb and fingers, except for the little finger. The aching that extends up the arm with this syndrome may be confusing, but the symptoms typically waken the patient at night or appear during gripping, i.e. driving, holding a newspaper, or knitting.

Since radiographic spondylosis is common and often asymptomatic (Fig. 1) 2307 no investigation can substitute for a careful history and examination. Electromyograms are unreliable since irrelevant or unimportant abnormalities are often shown.

Plain lateral radiographs, taken flexion and extension, are important in excluding other diseases as well as allowing assessment of canal size, the range of movement of the cervical spine and intervertebral joints, and any instability. Confirmation of spinal cord or nerve root compression is best obtained by cervical myelography.

Cervical spondylotic myelopathy is often helped by wearing a collar to restrict movement. Severely affected or rapidly worsening patients require surgery: the choice of operation should be tailored to the biomechanics encountered in the individual patient. There is evidence to suggest that a ‘decompressive laminectomy’ performed in a mobile neck may itself cause later neurological damage; the dura (or cord if the dura is left opened) becomes adherent to the cervical muscles and the normal physiological movement of the dura and cord is disrupted, producing abnormal traction forces on the cord. Decompressive laminectomy should be performed only on a patient with a diffusely narrow canal which is relatively immobile (less than 30° of head and neck movement between flexion and extension). More localized spondylosis is best treated by anterior cervical fusion with microsurgical removal of the intervertebral disc and replacement with an iliac bone graft. Many surgeons attempt to drill away osteophytes but this is probably unnecessary and may be dangerous: osteophytes are resorbed if the movement ceases. If indicated, laminectomy may follow anterior fusion, the latter procedure having produced a relatively immobile neck.

The results of such surgery are difficult to evaluate. In general, about 60 per cent of patients show neurological improvement, and the remainder cease to deteriorate. Total internal fixation of the head and neck, producing total immobilization, produces remarkable neurological improvement in elderly and bedridden patients, emphasizing the importance of movement in the pathogenesis of this condition and the importance of eradicating this movement in its successful treatment.

Cervical spondylotic radiculopathy usually subsides spontaneously, but in patients suffering severe or frequent episodes of pain, anterior cervical fusion may be performed with or without a bone graft, depending on the instability or otherwise of the joint. The results of fusion without removal of osteophytes are excellent in properly selected patients, over 95 per cent of whom are relieved of their root pain.

Acute cervical disc protrusion may occur spontaneously after a minor twisting injury or following a major flexion injury, such as diving into a shallow pool. This usually causes a radiculopathy that is often indistinguishable clinically and radiologically from osteophytic compression. Severe root pain is likely to be due to a disc protrusion, and this may coexist with an osteophyte. The treatment is the same, and the surgeon often sees a vertical rent in the posterior longitudinal ligament during microsurgical discectomy. Gentle exploration with a blunt hook allows the disc material to be teased out using the operating microscope.

More centrally placed disc protrusions cause spinal cord compression: a ‘partial Brown-Sequard syndrome’ may be seen if the protrusion is paracentral. Surgery using the anterior route allows the disc to be removed without difficulty. The results of surgery for cervical disc protrusions are, in general, excellent.

Lumbar disc protrusions are common, yet the results of surgery leave some patients dissatisfied. Removal of a lumbar disc protrusion relieves over 95 per cent of patients of their sciatica, but such results are only obtained if a definite disc protrusion compressing the nerve root is found and removed. Selection of appropriate patients is therefore the single most important factor determining the result. Bulging discs are normal and found in asymptomatic patients.

The usual history is that the patient bends and twists, and then suddenly feels his back ‘go’. He is locked in a flexed position with the disc prolapse compressing the pain-sensitive posterior longitudinal ligament. After a few hours or days further extension of the prolapse causes nerve root compression and the patient develops pain down the posterior or lateral aspect of the leg, radiating below the knee into the calf, ankle, or foot. Pins and needles and numbness develop in the appropriate dermatome. Physical signs comprise those of nerve root damage, either L5 or S1 (common levels for lumbar disc prolapse being L4/5 and L5/S1), limitation of straight leg raising on the ipsilateral side and pain on the affected side when raising the contralateral leg, and local tenderness and stiffness of the back, with scoliosis due to muscle spasm.

The most important signs indicating a lumbar disc protrusion are pain on raising the contralateral leg, calf wasting (measured with a tape measure), impairment or absence of the ankle reflex, weakness of dorsi flexion of the foot, and scoliosis of the lumbar spine.

The vast majority of patients recover with time; if this is to occur, it is obvious by 5 or 6 weeks from the onset. A patient seen at this time with a history of sciatica (rather than back pain) and with definite signs, it is reasonable to advise further investigation. Although the best method of investigation is water soluble myelography CT scanning has become fashionable. However, a considerable number of false-positive and false-negative results occur when CT scanning is undertaken without intrathecal contrast. CT scanning is best performed immediately after myelography when the myelographic abnormality is uncertain; scanning with contrast outlining the nerve roots is much more helpful.

MRI where available, has now replaced myelography as the investigation of choice. There is no place for discography in the investigation of sciatica.

Surgery should be advised if the patient can no longer cope with sciatic pain and if there is definite clinical and radiographic evidence of a disc protrusion compressing the nerve root. The usual indication is therefore pain, but occasionally a painless foot drop in a young active person may best be treated by removal of the prolapse, allowing the nerve root to recover. Back pain without significant sciatica should not, in general, be treated by removal of a disc protrusion, although a central disc protrusion (rather than a bulge) may occasionally cause back pain without sciatica. There is usually a history of trauma, and the patient finds sitting particularly difficult.

A central disc protrusion may also cause cauda equina compression, which is often associated with bilateral sciatica, perineal (pericoccygeal) numbness and tingling, and retention of urine. This is a surgical emergency requiring urgent myelography and removal of the disc prolapse; bladder, bowel and genital function will only recover following prompt and gentle surgical intervention.

The most important factor determining a good result is patient selection. Some surgeons advocate microsurgical removal, but this has no advantage, and the results are probably less good because adequate decompression of the nerve root is less likely using this restricted approach. There is also a danger of operating at the wrong level.

Insertion of an epidural catheter during surgery allows postoperative epidural diamorphine to be given. Encouraging patients to take adequate analgesia and keep moving allows mobilization within 48 h and discharge, on average, 3 or 4 days postoperatively. The patient is by then taking oral analgesics. Patients are instructed in back extension exercises: it is essential to re-educate the sacrospinalis muscles which are often weak, particularly if the patient has been resting in bed or using a corset.

Epidural fat grafts are pointless; every patient develops epidural scarring and this does not cause pain. Fusing the back in conjunction with removing a disc is wholly unnecessary. Percutaneous disc aspiration is inappropriate treatment for patients with a disc prolapse compressing a nerve. This is usually performed on patients with ‘bulging discs’ who should not be offered surgery anyway.

The most common complication is probably failure to cure the patient. If no disc protrusion compressing the nerve root is found the patient is likely to continue to have pain. Unfortunately a failed operation leaves the patient worse off, and they enter a group referred to as having a ‘failed back’. The failure in fact is that of the surgeon who was insufficiently careful about selecting the patient. Six weeks' relief of pain is common, even after inappropriate surgery: this appears to be due to a placebo effect. ‘Facet injections’ for back pain produce a similar pain-free period.

Disc prolapse may recur. The patient has clear-cut relief of sciatica (for longer than 6 weeks), followed by pain described as ‘exactly the same as before’. Proof of prolapse is not always clear on reinvestigation, due to postoperative scarring, and it may be necessary to explore the disc space.

Discitis may occur some days or weeks post operatively. This appears to be a low grade infection in the disc space producing back pain without sciatica. Radiographs or CT scanning show erosion of the cortical margins and the ESR is elevated. Bed rest is necessary, but the condition usually settles, often with spontaneous bony fusion of the disc space and excellent pain relief.

Peroperative nerve root damage should be very rare; the usual cause is inadequate exposure of the nerve root especially if it is tucked under the facet joint and is compressed by a lateral disc prolapse. Failure to expose and decompress the nerve root adequately leads to excessive root retraction.

A central disc prolapse also requires generous bony decompression prior to removal; attempts to remove a central protrusion through a fenestration (by just removing the ligamentum flavum) leads to severe and irreversible cauda equina damage, with disastrous consequences for the patient.

Lumbar spondylosis and facet joint degenerative hypertrophy may cause symptoms of nerve root compression. Two syndromes, one well-defined and the other less well-defined, occur.

Cauda equina claudication is a syndrome of sciatica with or without paraesthesia, which affects the patient when standing or walking but is completely relieved by sitting or lying down. The distance which can be walked gradually lessens with time; patients can walk further if they flex slightly and can usually bicycle without pain (which obviates the ischaemic theory of this condition). The symptoms can be bilateral or unilateral, and the only sign is that one or both ankle jerks are absent, especially if tested immediately after walking. Unlike patients with vascular insufficiency of the legs, patients with the cauda equina syndrome have to sit down to relieve pain; the former group can obtain relief by standing.

The usual cause of this syndrome is lumbar canal stenosis due to spondylosis superimposed on a congenitally narrow spinal canal. The ligamenta flava bulge forward in extension, hence the relative relief obtained by flexing the spine. Patients are usually over the age of 60; they may be imprisoned by pain, being unable to walk beyond the front door of their homes. Decompressive laminectomy produces excellent results, provided that the symptoms are exactly as described as above, the ankle jerk (or jerks) is absent, and MRI or myelography shows a complete or near complete block (Fig. 17) 2323. These may show a near complete block in asymptomatic individuals; the history and clinical findings are essential in making the diagnosis. A spinal angioma may mimic this syndrome and should be considered if there are doubts.

Younger patients with a disc protrusion may also present with symptoms of cauda equina claudication; it has been suggested that adoption of a vertical posture causes further protrusion to occur, but this seems a somewhat facile explanation.

This is a ill-defined syndrome of intermittent sciatic pain which is worse on standing and walking, but in a less well-defined and consistent manner than is seen in patients with cauda equina claudication. Appropriate nerve root signs are found and CT myelography reveals the nerve root trapped in the lateral gutter; that is between, anteriorly, an osteophyte and posteriorly, by a hypertrophic degenerative facet joint. There is, however, no appreciable canal stenosis (Fig. 18) 2324. This is a less well-defined syndrome, and the radiographic appearances are often seen in asymptomatic individuals. Decompression of the nerve root is justified in patients with sciatica (rather than back pain), with signs of nerve root damage, and with appropriate radiographic appearances. Surgery produces relief of sciatica pain in 60 to 70 per cent of patients.

Such surgery does not weaken the spine and by 3 months after surgery the patients can be encouraged to lead an entirely normal life, as long as back extension exercises are performed to counteract the mechanical abnormalities of the spine which arise from the previous disc prolapse or spondylosis.

This is rare, but important. The patient presents with progressive spinal cord compression, and often has a history of previous trauma or a fall from a height. Back pain may predate the onset of cord or root features. Thoracic disc protrusions are unlike protrusions elsewhere: they are often calcified and may erode through the dura. Very occasionally they are liquefied. The tumour-like progressive clinical picture may be a result of the immobility of the thoracic spine, due to its splintage by the ribs.

A thoracic disc prolapse is best demonstrated by CT myelography. Surgical treatment is required, since once neurological symptoms or signs develop progressive deterioration follows. The surgical approach to a central protrusion should be transthoracic. If the protrusion is laterally placed, a posterolateral approach removal of the transverse process and pedicle is satisfactory. A laminectomy should never be performed as this will almost certainly cause cord damage. Whichever approach is used great care is necessary to identify the correct level by X-ray marking of the rib or transverse process. The use of microsurgical techniques allows cord damage to be avoided.

The surgical relief of nerve root compression is beneficial to the patient with incapacitating sciatica. Back pain is not usually so amenable to surgery. Pain down the front of the thigh indicates a high lumbar root lesion (L1, 2, 3) and is not due to sciatic nerve or root compression (L4, 5, and S1). Referred pain from the back may extend down the back of the thigh to the knee, but pain extending beyond the knee, especially when associated with paraesthesiae is due to nerve root compression.

There are three clinically well-defined ‘sciatica syndromes’ (Table 3) 601 which should be considered while taking a history. Myelography (taken to the midthoracic area) is still a good method of investigating sciatica but MRI is preferable.

Back pain is not amenable to surgical treatment except when it is (rarely) caused by central lumbar disc protrusion, or when it occurs in association with spondylolisthesis (congenital as opposed to degenerative). This may cause severe low back pain in adolescents, and may require lumbar fusion. Back pain due to lumbar spondylosis is not generally relieved by fusion. The results are uncertain, but back pain which is completely relieved by bed rest might benefit from such treatment. The problem is two-fold: first, the origin of degenerative back pain is obscure; rigid anterior and posterior fusion often fails to relieve the pain. Second, radiography and MRI invariably show disc degeneration after the age of 40 years. The surgeon faced with a patient complaining of back pain may have to operate on radiographic appearances rather than on any clinical syndrome—a very unsatisfactory basis indeed. It is in the patient's best interest if the surgeon does not advise surgery for degenerative back pain.

Provocation discography, which involves injecting contrast medium into the disc space in an attempt to reproduce the pain, assumes incorrectly that back pain arises from the disc space. Clearly it does not, and this investigation should not be undertaken.

The vast majority of patients with pain following back surgery should never have been operated on in the first place. Patients with such a ‘failed back’ invariably suffer worsening symptoms and become crippled. Their lives (and those of their relatives) become totally dominated by the back pain, and compensation claims often exacerbate the situation. Unrealistic expectations of surgery must be corrected preoperatively. Although the patients' psychological characteristics must be considered, it is equally important to realize that long-standing pain, often causing sleep deprivation, is often associated with depression. There is a tendency to label patients with cauda equina tumours as hysterical or malingering, because their pain seems out of all proportion to their physical signs; it is (Fig. 21) 2327.

The patient's reaction to ‘back trouble’ may be analysed into three components. The first is the organic disease, perhaps the disc prolapse compressing the nerve. The second is the neurotic reaction, such as depression or anxiety. A certain amount is normal but in some patients this achieves undue proportions and, however effective the surgery to the nerve root, the anxiety or depression may, if excessive, mar the result. The third reaction is that of ‘illness behaviour’. This means abnormal behaviour in reaction to the disease, more explicitly a desire not to get better, because of either domestic or financial gain with illness behaviour; the patient gains either domestic attention which otherwise would not be available (pain becomes an old friend) or gains financially. In the latter situation the patient might act quite normally at home and reserve his symptoms for the assessing doctors. Illness behaviour in these circumstances has now become frank malingering.

The features of patients who maintain an illness behaviour need to be understood. They are remarkably well defined: pain is constant, present at all times, and spreads to other parts of the body. The pain gradually worsens, especially after a therapeutic intervention. Examination should start before the patient is seen: does he sit easily in the waiting room? How does he walk in? Does he sit without grimacing? Does he move like a patient in pain? Can he undress and get on and off the couch? A patient may produce a dramatic entry, and show great difficulty undressing or getting on and off the couch. However, while getting on the couch the patient may inadvertently sit at a right angle, demonstrating full straight leg raising, although this is envitably much reduced on formal testing. To any experienced clinician it is obvious that the patient does not behave like a patient with real pain. There is usually a marked lack of effort on testing motor power. Sensory loss is in a non-organic distribution, and usually extends up to the knee or thigh in a stocking distribution. Yet there are no definite neurological signs. The back is tender to the slightest touch and when the patient is asked to sit up to allow examination of the back this is (usually) done easily with relative enthusiasm to show the doctor the bad back, again demonstrating full straight leg raising. Pressure on the top of the head often produces a complaint of low back pain (which is not possible in mechanical terms).

If frank malingering is suspected it is often rewarding to observe the patient leaving the hospital.

Clearly surgery in these patients is doomed to fail, but it is surprising how many have been previously offered spinal fusion, suggesting illness behaviour is not as widely recognized as it should be.

Box CR. A case of invasion of the cauda equina by tumour with demarcation of all the sensory root areas of the lower limbs. Lancet, 1903; ii: 1566–72.

Fearnside MR, Adams CBT. Tumours of the cauda equina. J Neurol Neurosurg Psych, 1978; 41: 24–31.

Jayson MIV, ed. The Lumbar Spine and Backpain. 3rd edn. London: Churchill Livingstone, 1987.

Kerr RSC, Cadoux-Hudson TA, Adams CBT. The value of accurate clinical assessment in the surgical management of lumbar disc protrusion. Neurol Neurosurg Psych, 1988; 51: 169–73.

Tarlov E, D'Costa D. Back Attack. Boston: Little Brown and Co, 1985.

Vinken PJ, Bruyn GW eds. Handbook of Clinical Neurology. Volume 26 Part II. Injuries of Spine and Spinal Cord. Ch. 9. Cervical spondylotic radiculopathy and myelopathy. New York: American Elsevier Publishing Co: 97–112.