Back pain and leg weakness

Back pain is typically a recurrent condition. Although the prognosis for patients with an acute episode is favourable, with many recovering within 6–12 weeks, at least a third of such patients go on to have a recurrent episode within 1 year of the initial episode.5 Recurrent episodes contribute to much of the burden of back pain.

The major causes of back pain are shown in Box 1. The most common cause (> 85% of patients) is presumed musculo-ligamentous injury or “strain”, for which there is no clear pathophysiological basis and no identifying features on currently available clinical tests. It is most usefully classified as non-specific or idiopathic.3,4 Less common causes include radiculopathy (eg, disc herniation or spinal stenosis), referred visceral pain (eg, penetrating peptic ulcer), and other serious spinal conditions (eg, vertebral fracture, malignancy or osteomyelitis).

As non-specific back pain commonly resolves spontaneously, and given that serious aetiologies are rare, undertaking extensive investigations routinely is wasteful and potentially counterproductive. Conversely, history taking and examination should be done carefully to disclose any red flags (Box 2), which raise the potential of more serious underlying causes and demand a more intensive approach.1 In broad terms, these red flags relate to four diagnostic groups: serious neurological illness (eg, myelopathy, cauda equina syndrome), systemic illnesses (eg, malignancy, infection), vertebral fractures, and abnormal illness behaviour (eg, somatisation, psychogenic pain, malingering) (Box 2).

Information should be obtained about precipitating, exacerbating and relieving factors, and the location and radiation of the pain. Although pain may be described in mechanical or neuropathic terms, such terminology does not necessarily predict the underlying pathological process or the response to treatment. However, as illustrated in Hilary’s case, radiation of pain below the knee (“sciatica”), especially with accompanying numbness or paraesthesia, strongly suggests irritation of a lumbosacral nerve root. In these patients, leg pain that is exacerbated by twisting the back, sneezing or a Valsalva manoeuvre implicates herniation of a disc (Box 3), whereas exertional pain in the calf which is eased by forward flexion (eg, walking uphill or pushing a shopping trolley) may suggest spinal canal stenosis. If none of these classic histories is obtained, other potential causes of lumbosacral radiculopathy, plexopathy or sciatic neuropathy should be considered.

It is important to attempt to localise the neuroanatomical cause of the clinical signs, as multilevel disc herniations are common in older adults and it can be unclear radiologically which level is primarily responsible for a patient’s symptoms. Clinical and radiological correlation can confirm an L4/L5 disc protrusion — most commonly in the posterolateral direction, which can compress the L5 nerve root in the lateral recess of the spinal canal (Box 3), and less commonly in the far lateral direction, which can compress the descending L4 nerve root.

When the onset of back pain is historically remote from presentation with a neurological deficit, other neurological causes warrant exclusion, particularly in older patients who have multiple comorbidities (eg, disc herniation and motor neurone disease). Discomfort may limit a patient’s ability to cooperate with examination, but encouragement must be given to fully attempt motor tasks. It is often helpful to get the patient to stand one-legged from a chair (L3/L4), or stand on the heel (L5) or toes (S1) of one foot, to unmask subtle weakness that is not readily detectable on the examination couch.

The sciatic (L5/S1) and femoral (L3/L4) stretch tests are reasonably sensitive, but not specific for the presence of radiculopathy.

Sciatic stretch test: This test is commonly called the “straight-leg raising test” or Lasègue sign, and is performed with the patient in the supine position with the pelvis stabilised; the result is positive if calf pain is elicited when the leg is raised to less than 70 degrees (although lower limits are used in some studies).7 Straight-leg raising is a highly sensitive test for L5/S1 radiculopathies (positive in 90% of cases),7 but it is non-specific and may be positive in many other conditions (eg, hip osteoarthritis, trochanteric bursitis). Conversely, the “crossed straight-leg raising test” (where straight leg raising induces contralateral pain) is insensitive (positive in about 30% of cases) but highly specific for radiculopathy (about 90%).7

Femoral stretch test: This test is performed lying on the unaffected side, with the pelvis stabilised and the affected leg pulled posteriorly.

Trying to localise the site of a radiculopathy based on the motor signs is challenging, as most muscles are innervated by more than one nerve root and this varies between individuals. The overall pattern of weakness is more critical than the function of any single muscle being tested (Box 4). Consideration should be given to the pretest probability of finding an abnormality and the knowledge that 95% of all symptomatic disc herniations affect the L5 and S1 roots. Sensory symptoms are usually quite accurate in terms of localising the lesion, but sensory deficits may be ill defined or absent due to overlap of dermatomes (Box 4).

The differential diagnoses of a footdrop (from caudal to cranial) include: common peroneal and sciatic neuropathies, lower lumbosacral plexopathy, radiculopathy, hemi-conus or hemi-cord lesions, motor neurone disease, and parasagittal cortical or subcortical lesions. These can usually be distinguished by history taking and examination alone.

A spontaneous-onset lumbosacral trunk plexopathy, while distinctly unusual, can be clinically identical to an L5 radiculopathy, and may require nerve conduction studies for differentiation, if clinical suspicion of plexopathy is high.

Urgent neuroimaging is indicated in situations where important neurological symptoms or signs are present in a patient with low back pain. This includes patients with extensive deficits, which are usually due to compression of multiple nerve roots in the cauda equina, and produce symptoms of peri-anal (“saddle”) sensory loss, bowel and urogenital dysfunction and variable leg numbness and weakness.9 Prolapse of a lower lumbar central disc is the commonest cause of such a disturbance, but malignancy must be excluded. Although computed tomography (CT) is widely available and provides good delineation of bony architecture, magnetic resonance imaging (MRI) is the modality of choice as it provides greater contrast between soft tissue structures and does not use ionising radiation. In Australia, this is best facilitated by urgent referral to a specialist (neurologist, neurosurgeon or spinal surgeon).

Expert opinion is unanimous in recommending that imaging not be performed in the absence of red flags (Box 2).1,3,10 As well as the high cost, routine early imaging has not been shown to improve a patient’s level of pain, function or satisfaction, in the short or long term (Grade A evidence).11 As disc herniations are commonly found in asymptomatic individuals, the demonstration of possibly incidental abnormalities may reinforce or prompt abnormal illness behaviour. In cases of mild radicular features and a low suspicion of malignancy or infection, a trial of conservative therapy is recommended before undertaking neuroimaging (usually a CT scan before referral to a specialist), since many patients’ symptoms will resolve spontaneously over several weeks. Neuroimaging is required if surgery is being considered for the treatment of persistent symptoms or neurological deficits.

Nerve conduction studies are only indicated when the diagnosis is uncertain — for instance, when differentiating a low-lumbar plexopathy from an L5 or S1 radiculopathy. Sensory responses should be preserved in radiculopathy, as compression usually preserves the dorsal root ganglion and its distal (afferent) arm.

In other clinical settings, diagnostic studies that can justifiably be ordered in patients with back pain include plain x-rays and/or bone densitometry (for suspected vertebral compression fractures), tests for inflammatory markers (for possible infection), and investigations for referred causes of back pain (eg, ultrasound for suspected abdominal aneurysm).

In cases of acute-onset back pain and radicular symptoms, it is important to educate the patient. Non-interventional treatment is usually indicated as more than three-quarters of patients recover spontaneously within several weeks (Grade B evidence).12 Despite widespread use, there is inconsistent evidence supporting treatment of sciatica with non-steroidal anti-inflammatory drugs (NSAIDs), traction, physiotherapy or immunosuppressive agents (Grade B evidence). As bed rest does not appear to hasten recovery or improve symptoms,13 patients should generally be encouraged to mobilise. Although there is a paucity of data guiding the choice of specific analgesic medication (Grade C evidence),12 it seems reasonable to extrapolate from back pain guidelines14 and to commence with paracetamol (1 g four times a day) before escalating to an NSAID, then opioid analgesia (slow-release formulations of oxycodone are generally used), and (if pain lasts more than 3–4 weeks) a tricyclic antidepressant (Grade B evidence). The speed of escalation in this treatment algorithm should be determined by the degree of patient discomfort. As a general rule, conservative treatments should be encouraged for several weeks.

The approach to interventional treatment is highly individualised. It depends greatly on the response to conservative measures, the presence of radicular signs, and patient preferences. Some evidence indicates that transforaminal corticosteroid injection improves short- but not long-term outcomes,15 but this is a controversial treatment (Grade B evidence)12 that is best reserved for patients wishing to delay or avoid surgery (and should therefore be ordered by specialists). Although there is good evidence that early surgery (6–12 weeks) hastens recovery (Grade A evidence), it may not improve long-term (> 1 year) outcomes.12 Various surgical approaches are available (open discectomy, micro discectomy and micro-endoscopic discectomy and transforaminal discectomy), but as none has proven superiority, patient preferences play a role in choosing between them (Grade B evidence). Fusion procedures should rarely be performed. Structured rehabilitation hastens recovery after surgery (Grade B evidence).16

3 Herniated disc

---

Degeneration or trauma affecting the annulus fibrosus allows the nucleus pulposis to escape, usually posterolaterally, and compress the descending nerve root in its lateral recess (eg, L5 nerve root at L4/L5). A far lateral disc prolapse may compress the exiting nerve root at the neural foramen (eg, L4 nerve root at L4/L5).

Figure reproduced with permission from the Massachusetts Medical Society, copyright 2001.3

4 Patterns of clinical signs associated with specific nerve root lesions6,8

Nerve root	Distribution of potential weakness	Hyporeflexia	Sensory symptoms
L3	Hip flexion; hip adduction; knee extension	Adductor ± knee jerk	Low anteromedial thigh to knee
L4	Hip abduction and internal rotation; knee extension; foot dorsiflexion and inversion	Knee jerk	Knee and medial shin
L5	Foot and toe dorsiflexion; foot inversion and eversion; knee flexion; hip extension, abduction and internal rotation	Hamstring jerk	Anterolateral leg to hallux
S1	Ankle and toe plantar flexion; toe dorsiflexion; ankle eversion; knee flexion; hip extension	Ankle jerk	Posterior thigh and calf; lateral foot and sole

---

Approximate dermatomes for nerve roots in the leg are shown in the figure (bold lines indicate non-consecutive dermatomal boundaries, where overlap is less marked). Because of dermatomal overlap, radiation of sensory symptoms is a more reliable guide than sensory loss for localising the site of a radiculopathy.

---

Figure (dermatome diagram only) reproduced with permission from Elsevier, copyright 2010.6