Magnetic resonance imaging, which depends on the relaxation time of protons exposed to a magnetic field, demonstrates reliably the lesions of multiple sclerosis. Discrete, focal, or confluent areas of periventricular, callosal, pontine, medullary, or cerebellar white matter demyelination are seen in 98 per cent of patients. Spinal cord abnormalities occur in patients with normal cerebral scans and, unlike the cerebral lesions, these are rarely seen in normal individuals (Thorpe et al. 1993). Abnormalities can often be detected in the anterior and posterior visual pathways. The lesions correspond to areas of histological damage but are not specific for any one pathological process or diagnosis. However, characteristic patterns and distributions of proton density and T2-weighted lesions are seen which make the diagnosis of multiple sclerosis highly likely in the right context. These lesions are dynamic, and the element which fades or disappears is explained by alterations in water content due to resolution of oedema. Magnetization transfer imaging and T1-weighted sequences may be preferable for showing parenchymal destruction. Cerebral lesions are perhaps best demonstrated using fluid-attenuated inversion recovery (FLAIR) and this may be more specific for the spinal cord (reviewed by McDonald 1998b).
Gadolinium–DTPA crosses the blood–brain barrier when vascular permeability is increased and image enhancement indicates active inflammation. Serial studies of individual patients have established important principles about the dynamics of plaque formation. The earliest change seen in an evolving lesion is an increase in blood–brain barrier permeability; new lesions are first recognizable as areas of gadolinium enhancement which last for approximately 4 weeks and precede the onset of T2-weighted magnetic resonance changes or symptoms, by up to 2 weeks. As other features of the lesion develop, persistent enhancement is seen to occur as a ring around the edge of the lesion, although a uniform increase in signal also occurs. The lesions may recur in individual lesions and the cycles tend to complete within about 8 weeks. Enhancement is most obvious in the relapsing–remitting phase of the disease. It continues during secondary progression but is less evident in patients with primary progressive multiple sclerosis (Thompson et al. 1991).
All serial studies have shown that magnetic resonance lesions occur many times more frequently than new clinical events, especially with the use of triple dose delayed enhancing protocols. Some develop in areas which are strategically placed so as not readily to produce clinical symptoms but an alternative explanation is that many lesions never evolve to the stage at which pathophysiological disruption occurs. The dynamics of lesion formation are further complicated by studies showing that changes in magnetization transfer anticipate, by up to 3 months, areas which later enhance (Filippi et al. 1998). It has proved difficult to account for disability on the basis of conventional magnetic resonance imaging measurements, although lesion load in and around the corticospinal tracts matches scores on the EDSS (Riahi et al. 1998) and there is a much better correlation between disability and imaging abnormalities dependent on axonal loss than white matter involvement both in the cerebrum (De Stefano et al. 1998) and spinal cord (Stevenson et al. 1998).
Remarkably, areas of the contralateral cortex homologous to those involved in active demyelination show transient spectroscopic changes consistent with temporary axonal dysfunction, indicating widespread structural or metabolic effects of inflammatory demyelination in the central nervous system (De Stefano et al. 1999). Axonal loss produces atrophy, as may demyelination, but quantitative spectroscopic changes in N-acetyl aspartate are reliably associated with axonal loss both in affected and normal-appearing white matter (Fu et al. 1998).
Isolated enhancing lesions can be detected in most patients with optic neuritis (Youl et al. 1991; Gass et al. 1996). Long intracanalicular lesions are associated with poor visual outcome. Focal abnormalities, correlating with the clinical syndrome, can also be detected in patients with brainstem and spinal cord demyelination. Lesions which do not match the presenting clinical symptoms and signs are often present in patients with isolated demyelination. These are the harbingers of recurrent clinical activity. In a definitive prospective series, 85 per cent of those with additional lesions had developed clinically definite multiple sclerosis at 10 years compared with 11 per cent of patients with no other lesions (O'Riordan et al. 1998).
Cerebrospinal fluid analysis provides qualitatively different but complementary information in patients suspected of having multiple sclerosis. There is an increase in cell count, usually due to a lymphocytic pleocytosis rarely exceeding 50 cells/mm3, in about 50 per cent of patients and a modest rise in total protein, especially during periods of clinical activity. More sensitive and specific are increases in the immunoglobulin concentration and the presence of oligoclonal bands on protein electrophoresis, after correction for leakage of serum proteins through the blood–brain barrier. Both are evidence for intrathecal immunoglobulin synthesis and are seen in about 85 per cent of patients. Oligoclonal bands are seen in other diseases, many also associated with magnetic resonance abnormalities indistinguishable from the lesions of multiple sclerosis.
The specificity of the antibodies appearing in spinal fluid has not been resolved, but some at least are directed against components of the oligodendrocyte cell body or its myelin membranes and extrinsic viruses (Schadlich et al. 1987; Xiao et al. 1991). Molecular techniques provide an opportunity for identifying the antigens represented in these bands by screening against various libraries, and this approach has implicated Epstein–Barr virus nuclear antigen 1 in a small cohort of patients compared with controls (Rand et al. 1998). No treatments have reproducibly been shown to eliminate or reduce the number of oligoclonal bands in patients with multiple sclerosis.
Spinal fluid examination is particularly valuable in older patients who present some years after first developing symptoms, and in those with a late-onset progressive syndrome which may be due to demyelination—situations in which cerebral white matter lesions otherwise suggestive of multiple sclerosis can be age-related; the detection of oligoclonal bands is then highly informative and suggests inflammatory brain disease. The other situation where spinal fluid analysis can prove definitive is when spondylitic myelopathy has to be distinguished from primary progressive multiple sclerosis in patients with a progressive cord syndrome, equivocal changes of cord flattening on magnetic resonance imaging, and no cerebral white matter abnormalities. The presence of oligoclonal bands may then tilt the balance of probabilities in favour of demyelinating disease and spare the patient unrewarding spinal surgery.
Matthews (1998b) usefully summarizes the differential diagnosis of multiple sclerosis into the following categories: diseases that may cause multiple lesions of the central nervous system and also often follow a relapsing–remitting course; systematized diseases causing lesions in separate regions of the brain and spinal cord but usually with symmetrical manifestations and a progressive course; single lesions with either a remitting or progressive course; disorders which occur monophasically and at a single site; and non-organic symptoms which mimic the clinical manifestations and course of multiple sclerosis.
Clinical, immunological and imaging abnormalities indistinguishable from those of multiple sclerosis are seen in inflammatory disorders of the central nervous system. The cerebral or myelopathic form of systemic lupus erythematosus (Section 28.2) can occur in the relative absence of systemic manifestations and with only weakly positive serological abnormalities. Karussis et al. (1998) identified a group of patients with multiple sclerosis, many of whom had progressive myelopathy or optic neuropathy and an unusual clinical phenotype, in whom white matter lesions were associated with antiphospholipid antibodies. Primary Sjögren's syndrome can mimic multiple sclerosis, and there is evidence that these two conditions coexist more often than expected by chance. Sarcoidosis may present with widespread and relapsing involvement of the central nervous system, showing typical magnetic resonance and cerebrospinal fluid abnormalities, and in the absence of characteristic pulmonary or cutaneous manifestations (Zajicek et al. 1999). The distinction from multiple sclerosis cannot be made with confidence even in the presence of uveitis. A history of oro-genital ulceration in a patient with the clinical manifestations of multiple sclerosis should suggest the diagnosis of Behçet's disease (Section 28.11).
Racial differences in susceptibility make it necessary to consider alternative diagnoses when multiple sclerosis is diagnosed in individuals of African or Oriental origin. In both groups, the development of a progressive spinal disorder even with visual involvement is more probably due to HTLV-1 associated tropical spastic paraplegia. In Orientals, Devic's disease (see above) is more common than multiple sclerosis as a cause of spinal and visual pathway demyelination. Direct infection of the nervous system may mimic the syndromes of acute isolated demyelination or multiple sclerosis; these include tuberculous and other potentially chronic meningitides, the protean neurological manifestations of acquired immunodeficiency syndrome, and Lyme disease. The characteristic painful polyradiculitis and facial palsy that epitomize Borrelia infection in areas where this form of tick-borne spirochaetal infection has hitherto been uncommon, does not cause confusion, but the suggestion that borreliosis may produce a chronic or relapsing disorder of the central nervous system creates genuine diagnostic difficulty.
When considering systematized disorders, care must be taken in the diagnosis of multiple sclerosis if several affected members are identified in the same family. Pedigrees with hereditary spastic paraplegia (Section 14.8.2) mimic familial multiple sclerosis; and this may be the correct diagnosis in isolated cases of progressive spastic paraplegia, especially when the characteristic abnormality of gait occurs in the relative absence of muscle weakness. Other familial disorders that can be confused with multiple sclerosis include the hereditary ataxias and adult-onset leucodystrophies (Eldridge et al. 1984).
Although migraine, episodes of sudden onset indicating a vascular basis, early cognitive deficits, and a family history suggesting dominant inheritance do not immediately suggest multiple sclerosis, that is often the initial diagnosis in patients with CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy), especially when the cerebral white matter cerebral imaging appearances are seen in individuals with only ataxia and an unstable bladder (Dichgans et al. 1998); the distinction can be made by detection of the notch 3 gene mutation (encoded on chromosome 19p13) and by skin biopsy. Pedigrees characterized by affected males showing maternal inheritance may be examples of X-linked adrenoleucodystrophy (Moser 1997). Patients with clinically definite multiple sclerosis occur in families which otherwise manifest the clinical and genetic features of Leber's hereditary optic atrophy (Harding et al. 1992; Riordan-Eva et al. 1995). Differences in the age and clinical manifestations of subacute combined degeneration of the spinal cord should prevent confusion with multiple sclerosis, although focal relapsing spinal syndromes, often accompanied by Lhermitte's symptom, may occur in B12 deficiency.
Isolated syndromes related to multiple sclerosis are discussed above. The least forgivable error in the context of demyelinating disease is to accept the diagnosis of multiple sclerosis in patients with a progressive history in whom investigations have failed adequately to exclude a structural lesion—those at the foramen magnum being particularly well placed to confuse the unwary through appearing to produce independent spinal and brainstem symptoms. Errors also occur when the progressive or relapsing symptoms of brainstem and spinal arteriovenous malformations are mistaken for multiple sclerosis. Increased public awareness of multiple sclerosis and its manifestations leads many individuals with predominantly sensory symptoms or dizziness to seek neurological advice. Frequent but brief symptoms not accompanied by physical signs can usually be dismissed, but those who always ignore these complaints may occasionally be surprised by the findings of their more compliant colleagues. This is very different from the fabrication of spurious symptoms by individuals seeking the dignity of a neurological diagnosis in the setting of neurotic or psychiatric disease. Their management requires experience and firm handling. As in the whole of clinical neurology, a clear distinction has to be made between malingering and the tendency for any patient to exaggerate genuine manifestations of demyelinating disease in order to get their symptomatic message across to the busy practitioner or specialist.