Diagnostic procedures should not be used for preliminary screening, except perhaps in emergencies when a complete neurologic evaluation is impossible. Evidence uncovered during the history and physical examination should guide testing.
Lumbar puncture (spinal tap)
Lumbar puncture is used to evaluate intracranial pressure and CSF composition (see Table 1: Approach to the Neurologic Patient: Cerebrospinal Fluid Abnormalities in Various Disorders), to therapeutically reduce intracranial pressure (eg, pseudotumor), and to administer intrathecal drugs or a radiopaque dye for myelography.
Relative contraindications include
If papilledema or focal neurologic deficits are present, CT or MRI should be done before lumbar puncture to rule out presence of a mass that could precipitate transtentorial or cerebellar herniation.
For the procedure, the patient is typically in the left lateral decubitus position. A cooperative patient is asked to hug the knees and curl up as tightly as possible. Assistants may have to hold patients who cannot maintain this position, or the spine may be flexed better by having patients, particularly obese patients, sit on the side of the bed and lean over a bedside tray table. An area 20 cm in diameter is washed with iodine, then wiped with alcohol to remove the iodine and prevent its introduction into the subarachnoid space. A lumbar puncture needle with stylet is inserted into the L4-to-L5 interspace (the L4 spinous process is typically on a line between the posterior-superior iliac crests); the needle is aimed rostrally toward the patient's umbilicus and always kept parallel to the floor. Entrance into the subarachnoid space is usually accompanied by a discernible pop; the stylet is withdrawn to allow CSF to flow out. Opening pressure is measured with a manometer; 4 tubes are each filled with about 2 to 10 mL of CSF for testing. The puncture site is then covered with a sterile adhesive strip. A post–lumbar puncture headache (see Headache: Post–Lumbar Puncture and Other Low–Pressure Headaches) occurs in about 10% of patients.
Normal CSF is clear and colorless; ≥ 300 cells/μL produces cloudiness or turbidity. Bloody fluid may indicate a traumatic puncture (pushing the needle in too far, into the venous plexus along the anterior spinal canal) or subarachnoid hemorrhage. A traumatic puncture is distinguished by gradual clearing of the CSF between the 1st and 4th tubes (confirmed by decreasing RBC count), absence of xanthochromia (yellowish CSF due to lysed RBCs) in a centrifuged sample, and fresh, uncrenated RBCs. With intrinsic subarachnoid hemorrhage, the CSF remains uniformly bloody throughout collection; xanthochromia is often present if several hours have passed after ictus; and RBCs are usually older and crenated. Faintly yellow fluid may also be due to senile chromogens, severe jaundice, or increased protein (> 100 mg/dL).
Cell count and differential and glucose and protein levels aid in the diagnosis of many neurologic disorders (see Table 1: Approach to the Neurologic Patient: Cerebrospinal Fluid Abnormalities in Various Disorders). If infection is suspected, the centrifuged CSF sediment is stained for bacteria (Gram stain), for TB (acid-fast stain or immunofluorescence), and for Cryptococcus sp (India ink). Larger amounts of fluid (10 mL) improve the chances of detecting the pathogen, particularly acid-fast bacilli and certain fungi, in stains and cultures. In early meningococcal meningitis or severe leukopenia, CSF protein may be too low for bacterial adherence to the glass slide during Gram staining, producing a false-negative result. Mixing a drop of aseptic serum with CSF sediment prevents this problem. When hemorrhagic meningoencephalitis is suspected, a wet mount is used to search for amebas. Latex particle agglutination and coagglutination tests may allow rapid bacterial identification, especially when stains and cultures are negative (eg, in partially treated meningitis). CSF should be cultured aerobically and anaerobically and for acid-fast bacilli and fungi. Except for enteroviruses, viruses are seldom isolated from the CSF. Viral antibody panels are available. Venereal Disease Research Laboratories (VDRL) testing and cryptococcal antigen testing are often routinely done. PCR tests for herpes simplex virus and other CNS pathogens are increasingly available.
Normally, CSF:blood glucose ratio is about 0.6, and except in severe hypoglycemia, CSF glucose is typically > 50 mg/dL (> 2.78 mmol/L). Increased CSF protein (> 50 mg/dL) is a sensitive but nonspecific index of disease; protein increases to > 500 mg/dL in purulent meningitis, advanced TB meningitis, complete block by spinal cord tumor, or a bloody puncture. Special examinations for globulin (normally < 15%), oligoclonal banding, and myelin basic protein aid in diagnosis of a demyelinating disorder.
CT provides rapid, noninvasive imaging of the brain and skull. CT is superior to MRI in visualizing fine bone detail in (but not the contents of) the posterior fossa, base of the skull, and spinal canal. A radiopaque contrast agent helps detect brain tumors and abscesses. Noncontrast CT is used to rapidly detect acute hemorrhage and various gross structural changes without concern about contrast allergy or renal failure. With an intrathecal agent, CT can outline abnormalities encroaching on the brain stem, spinal cord, or spinal nerve roots (eg, meningeal carcinoma, herniated disk) and may detect a syrinx in the spinal cord. CT angiography using a contrast agent can show the cerebral blood vessels, obviating the need for MRI or angiography.
Adverse effects of contrast agents (see Principles of Radiologic Imaging: Radiographic Contrast Agents and Contrast Reactions) include allergic reactions and contrast nephropathy.
MRI provides better resolution of neural structures than CT. This difference is most significant clinically for visualizing cranial nerves, brain stem lesions, abnormalities of the posterior fossa, and the spinal cord; CT images of these regions are often marred by bony streak artifacts. Also, MRI is better for detecting demyelinating plaques, early infarction, subclinical brain edema, cerebral contusions, incipient transtentorial herniation, abnormalities of the craniocervical junction, and syringomyelia. MRI is especially valuable for identifying spinal abnormalities (eg, tumor, abscess) compressing the spinal cord and requiring emergency intervention.
MRI is contraindicated in patients who have had a pacemaker or cardiac or carotid stents for < 6 wk or who have ferromagnetic aneurysm clips or other metallic objects that may overheat or be displaced within the body by the intense magnetic field.
Visualization of inflammatory, demyelinated, and neoplastic lesions may require enhancement with IV paramagnetic contrast agents (eg, gadolinium). Although gadolinium is thought to be much safer than contrast agents used with CT, nephrogenic systemic fibrosis (nephrogenic fibrosing dermopathy) has been reported in patients with impaired renal function and acidosis.
There are several MRI techniques (see Principles of Radiologic Imaging: Magnetic Resonance Imaging); choice of technique depends on the specific tissue, location, and suspected disorder:
Magnetic resonance angiography (MRA) uses MRI with or without a contrast agent to show cerebral vessels and major arteries and their branches in the head and neck. Although MRA has not replaced cerebral angiography, it is used when cerebral angiography cannot be done (eg, because the patient refuses or has increased risk). As a check for stroke, MRA tends to exaggerate severity of arterial narrowing and thus does not usually miss occlusive disease of large arteries.
Magnetic resonance venography (MRV) uses MRI to show the major veins and dural sinuses of the cranium. MRV obviates the need for cerebral angiography in diagnosing cerebral venous thrombosis and is useful for monitoring thrombus resolution and guiding the duration of anticoagulation. Magnetic resonance spectroscopy can measure metabolites in the brain regionally to distinguish tumors from abscess or stroke.
Ultrasonography can be used at the bedside (usually in the neonatal ICU) to detect hemorrhage and hydrocephalus in children < 2 yr. CT has replaced echoencephalography in older children and adults.
Cerebral catheter angiography
X-rays taken after a radiopaque agent is injected via an intra-arterial catheter show individual cerebral arteries and venous structures of the brain. With digital data processing (digital subtraction angiography), small amounts of agent can produce high-resolution images. Cerebral angiography supplements CT and MRI in delineating the site and vascularity of intracranial lesions; it has been the gold standard for diagnosing stenotic or occluded arteries, congenitally absent vessels, aneurysms, and arteriovenous malformations. Vessels, as small as 0.1 mm, can be visualized. However, its use has decreased dramatically with the advent of MRA and CT angiography. It is still routinely used when cerebral vasculitis is suspected and when angiographic interventions (eg, angioplasty, stent placement, intra-arterial thrombolysis, aneurysm obliteration) may be necessary.
Duplex Doppler ultrasonography
This noninvasive procedure can assess dissection, stenosis, occlusion, and ulceration of the carotid bifurcation. It is safe and rapid, but it does not provide the detail of angiography. It is preferable to periorbital Doppler ultrasonography and oculoplethysmography for evaluating patients with carotid artery transient ischemic attacks and is useful for following an abnormality over time. Transcranial Doppler ultrasonography helps evaluate residual blood flow after brain death, vasospasm of the middle cerebral artery after subarachnoid hemorrhage, and vertebrobasilar stroke.
X-rays are taken after a radiopaque agent is injected into the subarachnoid space via lumbar puncture. MRI has replaced myelography for evaluation of intraspinal abnormalities, but CT myelography is still done when MRI is unavailable. Contraindications are the same as those for lumbar puncture. Myelography may exacerbate the effects of spinal cord compression, especially if too much fluid is removed too rapidly. Rarely, myelography results in inflammation of the arachnoid membranes around the spinal nerves (arachnoiditis), which may cause chronic pain and paresthesias in the lower back and extremities.
Electrodes are distributed over the brain to detect electrical changes associated with seizure disorders, sleep disorders, and metabolic or structural encephalopathies. Twenty electrodes are distributed symmetrically over the scalp. The normal awake EEG shows 8- to 12-Hz, 50-μV sinusoidal alpha waves that wax and wane over the occipital and parietal lobes and > 12-Hz, 10- to 20-μV beta waves frontally, interspersed with 4- to 7-Hz, 20- to 100-uV theta waves. The EEG is examined for asymmetries between the 2 hemispheres (suggesting a structural disorder), for excessive slowing (appearance of 1- to 4-Hz, 50- to 350-μV delta waves, as occurs in depressed consciousness, encephalopathy, and dementia), and for abnormal wave patterns.
Abnormal wave patterns may be nonspecific (eg, epileptiform sharp waves) or diagnostic (eg, 3-Hz spike and wave discharges for absence seizures, 1-Hz periodic sharp waves for Creutzfeldt-Jakob disease). The EEG is particularly useful for appraising episodic altered consciousness of uncertain etiology. If a seizure disorder is suspected and the routine EEG is normal, maneuvers that electrically activate the cortex (eg, hyperventilation, photic stimulation, sleep, sleep deprivation) can sometimes elicit evidence of a seizure disorder. Nasopharyngeal leads can sometimes detect a temporal lobe seizure focus when the EEG is otherwise uninformative. Continuous ambulatory monitoring of the EEG (with or without video monitoring) over 24 h can often determine whether fleeting memory lapses, subjective auras, or unusual episodic motor behavior is due to seizure activity. If clinicians need to determine whether an episode is a seizure or a psychiatric disorder, a video camera may be used to monitor the patient while EEG is done in the hospital. This technique (called video EEG) is also used before surgery to see what type of seizure results from an abnormality in a particular epileptogenic focus.
Measurement of evoked responses (potentials)
Visual, auditory, or tactile stimuli are used to activate corresponding areas of the cerebral cortex, resulting in focal cortical electrical activity. Ordinarily, these small potentials are lost in EEG background noise, but computer processing cancels out the noise to reveal a waveform. Latency, duration, and amplitude of the evoked responses indicate whether the tested sensory pathway is intact.
Evoked responses are particularly useful for detecting clinically inapparent deficits in a demyelinating disorder, appraising sensory systems in infants, substantiating deficits suspected to be histrionic, and following the subclinical course of disease. For example, visual evoked responses may detect unsuspected optic nerve damage caused by multiple sclerosis. When integrity of the brain stem is in question, brain stem auditory evoked responses is an objective test. Somatosensory evoked responses may pinpoint the physiologic disturbance when a structural disorder (eg, metastatic carcinoma that invades the plexus and spinal cord) affects multiple levels of the neuraxis. Somatosensory evoked responses can also help predict the prognosis of patients in a coma, particularly those with hypothermia, when the usual bedside indicators are unclear.
Electromyography and nerve conduction studies
When determining whether weakness is due to a nerve, muscle, or neuromuscular junction disorder is clinically difficult, these studies can identify the affected nerves and muscles.
In electromyography, a needle is inserted in a muscle, and electrical activity is recorded while the muscle is contracting and resting. Normally, resting muscle is electrically silent; with minimal contraction, action potentials of single motor units appear. As contraction increases, the number of potentials increases, forming an interference pattern. Denervated muscle fibers are recognized by increased activity with needle insertion and abnormal spontaneous activity (fibrillations and fasciculations); fewer motor units are recruited during contraction, producing a reduced interference pattern. Surviving axons branch to innervate adjacent muscle fibers, enlarging the motor unit and producing giant action potentials. In muscle disorders, individual fibers are affected without regard to their motor units; thus, amplitude of their potentials is diminished, but the interference pattern remains full.
In nerve conduction studies, a peripheral nerve is stimulated with electrical shocks at several points along its course to a muscle, and the time to initiation of contraction is recorded. The time an impulse takes to traverse a measured length of nerve determines conduction velocity. The time required to traverse the segment nearest the muscle is called distal latency. Similar measurements can be made for sensory nerves. Nerve conduction studies test large, myelinated nerves, not thinly myelinated or unmyelinated nerves. In neuropathy, conduction is often slowed, and the response pattern may show a dispersion of potentials due to unequal involvement of myelinated and unmyelinated axons. However, when neuropathies affect only small umyelinated or thinly myelinated fibers (or when weakness is due to a muscle disorder), results are typically normal. A nerve can be repeatedly stimulated to evaluate the neuromuscular junction for fatigability; eg, a progressive decremental response occurs in myasthenia gravis.
Nerve and muscle biopsy are usually done simultaneously. Nerve biopsy can help differentiate axonal from demyelinating polyneuropathies when other tests are inconclusive. A nerve supplying the affected area should be chosen. If polyneuropathy may be caused by vasculitis, the sample should include skin to increase the chances of finding a characteristic vascular abnormality. If the biopsy shows that nerve endings are lost, skin punch biopsy can help confirm small-fiber polyneuropathy. Muscle biopsy can help confirm myopathies.
Last full review/revision August 2012 by Michael C. Levin, MD