Review

Clinically isolated syndromes suggestive of multiple sclerosis,

part I: natural history, pathogenesis, diagnosis, and prognosis
Lancet Neurol 2005; 4: 281–88
David Miller, Frederik Barkhof, Xavier Montalban, Alan Thompson, Massimo Filippi
MS NMR Research Unit,
Institute of Neurology,
In 85% of young adults with multiple sclerosis (MS), onset is a subacute clinically isolated syndrome (CIS) of the University College London,
optic nerves, brainstem, or spinal cord. Methods of assessing the prognosis for patients who present with a CIS London, UK (D Miller FRCP,
have been sought, because only 30–70% of patients with a CIS develop MS. When clinically silent brain lesions A Thompson FRCP);
Department of
are seen on MRI, the likelihood of developing MS is high. MS can be diagnosed within 3 months of CIS
Neuroradiology, VU Medical
presentation with certain MRI and CSF criteria. Disability from MS is less likely in patients with a CIS of optic Centre, Amsterdam,
neuritis or sensory symptoms only, few or no MRI lesions, a long period to the first relapse, and no disability Netherlands (F Barkhof MD);
after the first 5 years. Development of more reliable prognostic markers will enable new treatments to be Department of
Neuroimmunology, Hospital
targeted for those who are most likely to benefit. We encourage continued clinical and laboratory assessment of
Vall d’Hebron, Barcelona, Spain
patients with a CIS. (X Montalban MD);
Neuroimaging Research Unit,
Multiple sclerosis (MS) is a common chronic disorder implications for the diagnosis and treatment of Scientific Institute and
University Ospedale San
of the CNS, characterised pathologically by areas of individual patients.
Raffaele, Milan, Italy
inflammatory demyelination that spread throughout (M Filippi MD)
the CNS over time. Although the clinical course is Clinical presentation Correspondence to:
highly variable, most patients eventually develop Review of a large database of patients with CISs found Prof D H Miller, NMR Research
severe neurological disability. In 85% of patients who that 21% presented with optic neuritis, 46% with long- Unit, Department of
Neuroinflammation, Institute of
later develop MS, clinical onset is with an acute or tract symptoms and signs, 10% with a brainstem
Neurology, University College
subacute episode of neurological disturbance due to a syndrome, and 23% with multifocal abnormalities.1 London, London WC1N 3BG, UK
single white-matter lesion. This presentation is known Therefore, the syndrome was isolated in space in only d.miller@ion.ucl.ac.uk
as a clinically isolated syndrome (CIS). 77% of presentations, although all presentations were
When patients present with a CIS (eg, optic neuritis, isolated in time. The presentation of MS affects disease
or an isolated brainstem or partial spinal-cord course and prognosis (panel 1).2–15
syndrome), clinicians are faced with many questions. Studies of the clinical features and natural history of
Is the CIS due to a disorder other than MS? What is the CISs use many different groups of patients, either from
likelihood that the person will develop MS? If MS does natural-history studies or from short-term observational
develop, is it possible to ascertain the likelihood of or therapeutic studies. These two sources have
disability? How should the patient be assessed— complementary strengths and weaknesses. The natural-
with MRI, CSF analysis, or other tests? What should history studies are representative and provide good long-
patients be told? Should patients be treated to assist term data, whereas in short-term treatment studies, the
recovery from the CIS, or to delay the development data are closely sampled and highly reliable within the
of MS? follow-up period.
Groups of patients with a CIS have been the focus of Data on outcomes in untreated CIS and MS are
much published research, especially in recent years. available from three trials—North American Optic
There has been interest in MRI as a tool to assist with Neuritis Treatment Trial (ONTT);16 Controlled High
diagnosis and prognosis. The advent of disease- Risk Subjects Avonex Multiple Sclerosis Prevention
modifying treatments for MS has increased attention Study (CHAMPS);17 and Early Treatment of Multiple
on early disease. In some patients these treatments Sclerosis study (ETOMS).4 Comparison of outcomes in
might prevent the development of disability but in the placebo arms of these studies is not straightforward
others they could be superfluous. Because a CIS is because of differences in the clinical and MRI selection
typically the earliest clinical expression of MS, research criteria. The probability of a second episode during the
on patients with a CIS may provide new insights into study period was 16·7% in ONTT, 38% in CHAMPS,
early pathological changes and pathogenetic
mechanisms that might affect the course of the
Panel 1: Features of CIS and early MS reported to affect prognosis
disorder.
Our review of CISs is presented in two parts. In part Good prognosis Poor prognosis
1, we discuss the natural history (particularly in Optic neuritis2,3 “Multifocal” CIS4
relation to MS), pathogenesis (pathological and Isolated sensory symptoms5 Efferent systems affected5
immunological), diagnosis, and prognosis. In part 2, Long interval to second relapse2 High relapse rate in the first 2–5 years6
in the next issue (June 2005), we will cover No disability after 5 years2,5,7,8 Substantial disability after 5 years5
findings from non-conventional magnetic resonance Normal MRI9,10,11,12 Abnormal MRI with large lesion load9–15
techniques, recovery processes, and management—eg,

http://neurology.thelancet.com Vol 4 May 2005 281

 Review

and 45% in ETOMS. Prognosis seems to be poor with a involvement at initial presentation. They also proposed
multifocal presentation. In ETOMS, the chance of that although incomplete remission, multifocal
developing clinically definite MS was greater in patients presentation, and efferent involvement at presentation
with a polysymptomatic onset than in those with a were all predictive of poor outcome, most long-term
unifocal onset (odds ratio=1·99, 95% CI 1·14–3·46, deficit is secondary progression, which is not associated
p=0·015). No difference between the various symptoms with these early events.
was evident at onset. Earlier studies from London (Ontario, Canada) have
The presence and number of lesions on MRI has an reported a relation between relapse during the first
effect on the course of the disorder; the risk of having a 2 years and disability status after 11 years;6 the results of
diagnosis-defining second episode is higher in those a 25 year follow-up of this group have not been
with an abnormal scan.9,11,12,14,15,18 In a natural-history published yet.
study of a group of patients in London (UK) the Optic neuritis is the best-studied CIS. One of the
percentage who developed definite MS (including those longest follow-up studies of patients with optic neuritis
with and without abnormality on initial brain MRI) was was done at the University of Lund (Sweden) where 86
43% at 5 years, 59% at 10 years, and 68% at 14 years.10 patients with acute monosymptomatic unilateral optic
The presenting symptoms did not seem to affect the rate neuritis were recruited between 1969 and 1981 and
of conversion to definite MS, but lesions on MRI at followed up until a diagnosis of MS was made.3 Of the 86
baseline were strongly predictive of diagnosis, and the patients, 33 developed MS, three died, and 50 continued
number of lesions was related to time to relapse14 and to have isolated optic neuritis. 43 of the 86 patients were
subsequent disability.10,19 assessed: CSF was analysed at onset and MRI done
Another natural-history study, of 1215 patients with during follow-up. In this group, the estimated 15 year
MS in Lyon (France)2 found a median time to second risk of MS was 40% (95% CI 31–52%). In 60% of
episode of 1·9 years. In this study, there was no patients, MS occurred within 3 years. Of the factors
difference in the predictive value of unifocal or present at onset, those patients who had signs of
multifocal presentation, but there was a longer period to inflammation in the CSF (raised cell count, or
the second episode with optic neuritis presentation than oligoclonal bands, or both) had a 49% (38–65%) risk of
with either brainstem or spinal-cord presentations. Slow MS compared with those who had no CSF signs (23%
progression to disability status scale scores 4 and 6 also [12–44%], p=0·02). Similarly, recurrence of optic
occurred in patients with optic neuritis at presentation, neuritis was associated with a high risk of developing
in those with good recovery from the CIS, in those with a MS (p<0·001). Patients who were young and those with
long period before the second relapse, and in those with onset in winter also had a high risk of MS. In 30 patients
few relapses in the first 5 years.2 Overall, however, the with isolated optic neuritis, MRI 19–31 years after onset
CIS characteristics had little effect on delay to the second showed lesions suggestive of demyelinating disease in
episode and little, if any, effect on subsequent 20 of the patients, even though no clinical signs of MS
progression. had occurred.
Long-term follow-up of patients with a CIS from a Despite presentation with optic neuritis or brainstem
Gothenburg (Sweden) database found some overlap or spinal-cord syndromes, patients with a CIS may have
with those in the Lyon study, but also some differences.5 cognitive deficits. Overall scores on a composite battery
Study of 308 patients from the Gothenburg database of tests were low in a group of 48 patients with a CIS
found that only 45 of 220 patients with a CIS still had a compared with controls.20 When this group was followed
CIS at the 25 year follow-up. Patients with efferent up 4·5 years later, if they had no further episodes or a
lesions had twice the risk of a later diagnosis of MS than relapsing-remitting course, cognitive impairment was
those without efferent lesions, and the likelihood of not changed.21 However, those who developed
being wheelchair bound (disability status scale score 7) secondary-progressive MS had deteriorated on tests of
was 2·8 times higher with efferent than non-efferent attention and were also beginning to show signs of
lesions and 1·9 times higher with incomplete than memory deficits. More recent studies, albeit in early
complete remission of the CIS. Incomplete remission relapsing-remitting MS with magnetisation-transfer-
(3·1 times higher risk) and high relapse rate (2·8 times ratio MRI, suggest that cognitive impairment is related
higher risk) in the first 5 years of MS were predictive of to atrophy and abnormalities in normal-appearing brain
subsequent disability status scale score 7. However, the tissue.22,23 Magnetic resonance spectroscopy shows that
prediction from the relapse rate was valid only with a cognitive deficits relate to concentrations of N-acetyl-
cut-off at high rates of occurrence (five or more relapses aspartate in the brainstem.24,25
in 5 years). Indeed, the crucial feature for assessing the
eventual severity of disability—the rate of secondary Pathogenesis
progression—was not predictable from the early phase Pathology
of the disease. The researchers concluded that MS seems There are few, if any, studies on the pathological
to have a long preclinical history with widespread changes that occur in CISs, and most that there are

282 http://neurology.thelancet.com Vol 4 May 2005

 Review

focus on more aggressive forms of inflammatory than controls, and the number of cells detected is
demyelination such as Balo’s concentric sclerosis, associated with disease activity.33 Myelin-oligodendrocyte
Marburg’s disease, Schilder’s disease, and acute -glycoprotein (MOG) antibodies and MBP antibodies
disseminated encephalomyelitis. Initial presentation were measured in serum samples from 103 patients
with a large mass lesion (so-called tumefactive MS) may with a CIS, oligoclonal bands in their CSF, and positive
also pose diagnostic difficulties. However, the finding of MRI scans.34 A disease-defining relapse was seen in nine
many infiltrating macrophages in the setting of myelin (23%) of 39 seronegative patients, with a mean time to
loss with relative preservation of axons provides strong relapse of 45 months. A second episode occurred after a
support for inflammatory-demyelinating disease in mean of 7 months in 21 (95%) of 22 patients positive for
these patients. both antibodies and after a mean of 15 months in 35
An important pathological study in early MS showed (83%) of 42 patients positive for MOG antibodies.
heterogeneity in actively demyelinating lesions.26 Compared with seronegative patients, the adjusted
Although all had inflammatory infiltrates of hazard ratio for MS was 76·5 for patients who were
T lymphocytes, they segregated into four different seropositive for both antibodies and 31·6 for patients
patterns according to the distribution of myelin loss, the who were seropositive for MOG antibodies. Another
plaque geography and extension, the pattern of recent study reported high serum concentrations of
oligodendrocyte destruction, and immunoglobulin and MOG antibodies in patients with a CIS compared with
complement deposition. The authors suggest that any controls.35 Replication of these studies is needed and
single patient has only one of these four patterns. excitement over the results must be tempered by a
More recently Barrett and Prineas27 have highlighted recent study showing that patients with MS and healthy
questions fundamental to the pathology of MS and people have a similar proportion of anti-MOG IgG and
shown extensive oligodendrocyte apoptosis and IgM,36 and another study in which myelin antibodies in
microglial activation in myelinated tissue containing few patients with a CIS was not associated with either the
or no lymphocytes or myelin phagocytes. Such findings presence of abnormalities on MRI or the early
invite a reassessment of the pathology of MS, which may development of MS.37
well have implications for our understanding of the Detailed studies in patients with CIS suggest there
progression and treatment of the disorder.28 may be systemic activation of myelin-reactive T cells,
An old study reported that 59% of lesions in early, which secrete proinflammatory Th-1 cytokines.38 This
severe MS occured in grey matter,29 and recent activation has been associated with inflammatory activity
pathological work has focused on this region.30,31 Studies on gadolinium-enhanced MRI. Some research suggests
of grey matter have focused on progressive MS or on that regulatory T cells might be protective and a recent
early relapsing-remitting disease rather than on CIS. study has suggested a negative association between the
The most common lesions are intracortical rather than percentage of CD25+ CD4 T cells in the CSF and disease
full thickness, and lesions on the border of grey and activity.38 The chemokine receptor CXCR3 seems to be
white matter are also seen. Fewer macrophages are seen expressed in most CSF T cells in patients with CIS in
in grey-matter lesions but there are high numbers of whom the CSF concentrations of the CXCR3 ligand,
microglia. In intracortical lesions, CD3 cells are CXCL10 (IP-10), are selectively increased.39 However, a
completely absent, whereas purely cortical lesions are broad range of chemokines and chemokine receptors are
largely non-inflammatory with good preservation of the also expressed in the CSF in early MS.
cytoarchitecture. In summary, although there are some potential
In summary, the many pathological abnormalities in candidates, we do not have a validated immunological
both white and grey matter in early MS, in conjunction marker to predict the development of MS in patients
with magnetic resonance studies (see later), is with CIS.40 None of the immunological changes
compatible with similar pathological abnormalities described are specific to MS.
being present, albeit to a lesser degree, in patients with
CIS. Diagnosis
Clinical suspicion of demyelination is high when a
Immunology young person has an episode consistent with damage to
Studies of the immune activation in patients with CIS white-matter tracts; however, many diseases cause
provide important information about immunological similar episodes. A comprehensive review of the
mechanisms. An abnormal B-cell response (shown by differential diagnosis of MS is beyond the scope of this
IgG oligoclonal bands in the CSF) is detected in about review; however a few caveats should be mentioned.
two-thirds of patients with CIS, is best documented in First, the diagnosis in patients with CIS—MS, another
patients with optic neuritis, and increases the risk of neurological disorder, or CIS only—is a clinical decision
MS.32 In addition, cells in the CSF secreting antibodies and should be made by a neurologist experienced in the
for myelin-basic-protein (MBP) and phospholipid- relevant differential diagnoses. Certain inflammatory or
protein antibodies are found in more patients with CIS infectious disorders (eg, systemic lupus erythematosus

http://neurology.thelancet.com Vol 4 May 2005 283

 Review

and neuroborreliosis) that are commonly part of the Diagnostic criteria

differential diagnosis can be investigated through blood Several criteria have been proposed for the use of MRI as
and CSF analyses. MRI is crucial investigation for a diagnostic method. In 1988, Paty and colleagues47
differentiation of brainstem and spinal-cord CIS owing suggested the presence of four or more lesions, or three
to inflammatory demyelination from other structural lesions, of which one is periventricular as diagnostic of
lesions (eg, brainstem vascular malformation and cord MS. These criteria have been studied prospectively in
compression). Last, in patients with suspected optic CIS, showing high sensitivity but relatively low
neuritis, there are many clinical features that suggest a specificity.48 Fazekas and co-workers’49 widely used
cause other than inflammatory demyelination criteria require the presence of three or more lesions
(panel 2).41 with at least two of the following characteristics: an
By definition, patients who present with CIS do not infratentorial lesion, a periventricular lesion, or a lesion
have MS, though they are at a high risk. It might be larger than 6 mm. Retrospective study of these criteria in
important for the patient to know their likelihood of patients with established MS has shown both high
developing MS after one episode. As happens in many sensitivity and high specificity;50 however, these criteria
neurological disorders in which pathological material is are less useful in prospective studies or when applied to
rarely analysed and no current genetic testing is patients who present with a CIS.51
available for diagnosis, this must be done with clinical In 1997, Barkhof and colleagues46 developed criteria
and paraclinical data. Historically, the demonstration of for MRI assessment to improve the prediction of CIS
clinical dissemination in space and time of an developing into clinically definite MS. These were based
inflammatory-demyelinating process has been, in the on a cumulative chance model of four criteria: presence
absence of other diseases, the gold standard for the of at least one juxtacortical lesion, at least one
diagnosis of MS. The Poser committee criteria gained gadolinium-enhanced lesion, at least one infratentorial
widespread acceptance; these criteria incorporated lesion, and three or more periventricular lesions. The
information from laboratory investigations and allowed presence of at least three of the four features was shown
an early diagnosis within the category of laboratory- to be a more accurate predictor than the Paty or Fazekas
supported definite MS in those patients with an criteria for the development of MS.46 The Barkhof
abnormal result on CSF analysis.42 Abnormal intrathecal criteria were later modified by Tintoré and colleagues,52
IgG synthesis, defined as two or more oligoclonal bands who allowed nine T2 lesions in place of a gadolinium
in the CSF without corresponding bands in the serum, is enhancing lesion; the modified criteria were also
found in 60–70% of patients with CIS and is associated predictive in patients with CIS.52
with a high risk of developing MS. Early MRI studies In 2001, a panel of international specialists published
showed that 50–70% of patients with a CIS already had new guidelines for the diagnosis of MS.53 Like the Poser
T2 abnormalities in the cerebral white matter.43–45 Several criteria, these guidelines rely on objective evidence of
follow-up studies later showed that patients with lesions dissemination in time and space. In addition, they
on MRI commonly develop clinically definite accept that specific MRI abnormalities can provide this
MS,9–12,14,15,18,46 whereas patients without abnormalities on objective evidence in patients with CIS (figure). The
MRI have a better prognosis and most do not develop McDonald guidelines accept the modified Barkhof-
clinically definite MS, at least for the duration of Tintoré criteria for evidence of dissemination in space. If
available studies (up to 14 years). these more stringent imaging criteria are not fulfilled,
the presence of at least two T2 lesions plus the presence
of oligoclonal bands gives other evidence for
Panel 2: Warning signs in optic neuritis* dissemination in space. However, this alternative
criterion may decrease accuracy in the diagnosis: Tintoré
Optic atrophy on presentation
and co-workers54 reported that the criteria had a
Severe disc oedema with vitreous reaction
specificity of only 63% for the development of clinically
Optic-disc haemorrhage
definite MS after follow-up of patients with CIS for
Bilateral loss of vision
3 years. The MRI criteria for dissemination in time
Previous history of neoplasia
require the presence of a gadolinium-enhanced lesion
Afro-Caribbean ethnicity with vision 6/12 and no early recovery
on an MRI done at least 3 months after the first episode,
Loss of vision to no perception of light with no early recovery
or a new T2 lesion confirmed on two MRI scans, done
Painless loss of vision to 6/60 with no early recovery
3 months after the CIS.
Severe or persistent pain for 2 weeks since onset
Two studies have assessed the accuracy of the
Visual loss progressing for 2 weeks since onset
McDonald criteria to predict after 1 year whether
Lack of recovery 3 weeks after onset of visual symptoms
patients with CIS will develop clinically definite MS
Worsening of vision after withdrawal of corticosteroids
within 3 years. Dalton and colleagues55 found a
*These signs are atypical in optic neuritis and should alert clinicians to diagnoses other than inflammatory demyelination. sensitivity, specificity, and accuracy of 83%. Tintoré and
co-workers54 reported a sensitivity of 74%, a specificity of

284 http://neurology.thelancet.com Vol 4 May 2005

 Review

A B C D E

Figure: MRI scans of a patient with optic neuritis at presentation and 3 months later
At baseline (A, B, and C; all T2-weighted MRI), the McDonald criteria for dissemination in space are fulfilled, with more than nine T2 lesions, three periventricular lesions, and infratentorial and
juxtacortical lesions. After 3 months (D [T2-weighted MRI] and E [gadolinium-enhanced T1-weighted MRI]), there is dissemination in time with two new enhancing lesions (E), also seen as new
abnormalities on the T2-weighted scan (D).

86%, and an accuracy of 80%. When a new T2 lesion was cord lesions are rare in disorders other than MS, and an
allowed as evidence for dissemination in time, Dalton abnormal spinal cord was identified from MRI scans in
and colleagues56 showed that 14 (82%) of 17 patients who 86 (83%) of 104 patients who were recently diagnosed
fulfilled the new criteria for MS after 3 months had with MS and helped to show dissemination in space at
developed clinically definite MS after 3 years, compared the time of diagnosis.64 However, spinal-cord imaging in
with only five (13%) of 39 patients who did not fulfil the a group of 115 patients with optic neuritis did not affect
criteria. Similarly, Tintoré and co-workers54 found that subsequent diagnosis when the McDonald criteria, in
whereas 28 (80%) of 35 patients who fulfilled the new which one spinal-cord lesion can be substituted for a
criteria for MS after 1 year developed clinically definite brain lesion, were applied.65
MS after 3 years, only 10 (20%) of 51 who did not fulfil
the criteria did so. In addition, in the placebo arm of the MRI assessment of long-term prognosis
ETOMS trial, the McDonald criteria for dissemination in Although many studies assess the predictive value of
space were associated with an increased likelihood of MRI in the short term, few have addressed the mid-term
developing clinically definite MS.57 Though the initial to long-term, partly because this technique was
application of the McDonald criteria is favourable and introduced in the mid 1980s. Research into the
supports their use, there is scope for improvement. In a predictive value of MRI in the mid-term to long-term is
recent review by some of us,58 we discuss the pros and crucial. Too short a follow-up will lead to ascertainment
cons of the criteria and suggest an approach to their bias and might lead to overestimation of the role of MRI
application in patients with a CIS; we also propose that by identifying only those patients in whom the period for
dissemination in time be supported by a new T2 lesion conversion is short. There is also little change in
after 3 months instead of gadolinium-enhanced lesions. disability in short-term studies, which impedes
In a CIS of the brainstem, the dissemination in space investigation of the relation between disability and MRI.
criteria are not as specific in predicting conversion to Five studies—four published (table 1)10–12,66 and one
clinically definite MS as when they are applied to other unpublished (J Frederiksen, personal communi-
CISs; this is mostly because the presence of an cation)—with MRI findings from groups of patients
infratentorial lesion is not used to show dissemination with CIS have included a follow-up interval of at least 5
in space in this subgroup of patients.59 There are several years. The populations studied ranged from only optic
prospective studies of the clinical and MRI neuritis to mixed populations.10,12,66 A summary of the
characteristics of spinal-cord syndromes in patients with patients studied in the published research is given in
a first episode of MS.60,61 Patients typically present with table 1; the fifth study was of 183 consecutively referred
partial myelitis, and asymmetric clinical findings are patients with acute monosymptomatic optic neuritis
characteristic, commonly with predominant sensory
symptoms. Severe cases of acute transverse myelitis with
loss of all leg movements and spinal shock are
Study CIS type Number Mean duration of Patients who develop
uncommon. As in established MS,62 spinal lesions seen of patients follow-up (years) clinically definite MS (%)
on MRI typically extend over less than two spinal
Brex et al 200210 Mixed 71 14 68
segments.63 In all CISs, abnormal brain MRI is the Optic Neuritis Study Group 200311 Optic neuritis 388 10 38
strongest predictor of conversion, followed by the Minneboo et al 200412 Mixed 42 8 62
presence of oligoclonal bands. Nevertheless, the Tintore et al 200466 Mixed 136 6·5 46
potential for MRI of the spinal cord to aid diagnosis has Table 1: Baseline characteristics in long-term follow-up studies of CIS
recently gained much interest. Asymptomatic spinal-

http://neurology.thelancet.com Vol 4 May 2005 285

 Review

from Copenhagen, of whom 40% had developed

Normal MRI Abnormal MRI
clinically definite MS after a mean of 5 years.
Brex et al10 0% 45%
In general, patients with optic neuritis seem to have a Minneboo et al12 12% 39%
lower risk for conversion to clinically definite MS than Optic Neuritis Study Group69* 29% 38%
patients with another CIS, even after 10 years, probably
*These figures include only the subgroup of patients who developed clinically definite MS.
because—in some series—the incidence of MRI lesions
is lower for optic neuritis than for brainstem or spinal- Table 3: Deterioration to EDSS score 3, stratified by baseline MRI
cord syndromes.11,18,60 By contrast, the conversion rate is
higher in mixed populations, but does not approach but also that some new lesions will result from small-
100%, even after moderately long follow-up, suggesting vessel disease rather than demyelination when there is a
that even when demyelinating events happen outside the long period of follow-up.
optic nerve they can stay monophasic and that there is a Though it is important to predict the conversion to
need for paraclinical data to predict dissemination in clinically definite MS, prediction of the development of
space. The predictive value for conversion to clinically disability is perhaps even more important. The expanded
definite MS of the four published studies is provided in disability status scale (EDSS) has typically been used to
table 2;10–12,66 in the Copenhagen study, 60% with an measure disability and data on the relation between MRI
abnormal scan and 10% with a normal scan developed and EDSS scores are available.10,12,66,69 Table 3 shows the
clinically definite MS after a mean of 5 years percentage of patients who reach EDSS score 3
(J Frederiksen, personal communication). As in (representing mild, but definite disability) as a function
previous studies of shorter duration, MRI findings at of findings on baseline MRI in three of the four studies.
baseline are related to the risk of conversion. In patients The predictive value is consistently evident in two
with an abnormal scan who were followed up for a long studies that included a total of 113 patients from mixed
period, the conversion rate was more than 50% in onset populations with median follow-up of 8 years and
patients with optic neuritis and about 90% in patients 14 years.10,12 In the third study, the absence of a clear
with another CIS, although with little indication that the relation between MRI and disability is in striking
number of lesions increases the already high risk.10 109 contrast to the clear trends for conversion to clinically
patients with CISs were studied with MRI at baseline definite MS.11,69 The difference is partly explained by the
and 1 year;60,67 89 patients were monitored for 5 years,14 inclusion of only patients with MS in the analysis of
81 patients for 10 years,19 and 71 patients for 14 years.10 disability at follow-up, whereas all patients, including
After 14 years, patients with an abnormal MRI at those with CIS only, were included in the other two
baseline had a very high conversion to clinically definite studies. Since there are more patients with CIS only in
MS, independent of the overall severity of lesions (low the group with normal scans, their inclusion will lower
[89%] vs medium [87%] vs high [88%]).10 Several studies the disability of this group compared with those with
in other groups of patients with CIS have been abnormal scans.69 A fourth study of a mixed CIS
reviewed.68 population supports the link between MRI and
When MRI does not show signs of abnormality, disability: the number of lesions at baseline and EDSS
conversion to clinically definite MS is unlikely but scores after a mean of 6·5 years are correlated (r=0·4,
possible; this could be explained by the CIS attack being p<0·001).66
the first instance of demyelination. After 14 years follow- The number of lesions at baseline was related to the
up of the London (UK) group, 19% of the patients with a risk of reaching EDSS scores 3 and 6 after 14 years
normal MRI at baseline developed clinically definite MS (moderate disability; table 4).10 Analysis of lesion loads at
and an additional 21% developed new MRI lesions.10 years 5, 10, and 14 of that study showed that the
Long-term follow-up studies are likely to be affected by strongest correlation of final disability was with the
dropouts. Also, imaging quality and resolution are likely measures at 5 years, and that subsequent changes in
to have improved during the period of study: not only is lesion load are less relevant.10 Pathogenetic processes
it possible that small lesions were not always detected in additional to T2-visible lesions (in normal-appearing
some patients in the initial study because MRI white and grey matter) might become more important
resolution was not as good as that at the last follow-up, over time.

Number of lesions on MRI at baseline

Study Normal MRI Abnormal MRI
0 1–3 4–10 10
Brex et al
10
19% 88%
Optic Neuritis Study Group11 22% 56% EDSS 3 0% 31% 54% 80%
Minneboo et al12 24% 72% EDSS 5·5 0% 13% 31% 60%
Tintore et al66 8% 63% EDSS 10 0% 0% 8% 13%

Table 2: Development of clinically definite MS, stratified for baseline Table 4: Relation between number of lesions at baseline and later EDSS
MRI findings score10

286 http://neurology.thelancet.com Vol 4 May 2005

 Review

2 Confavreux C, Vukusic S, Adeleine P. Early clinical predictors and

Search strategy and selection criteria progression of irreversible disability in multiple sclerosis:
an amnesic process. Brain 2003; 126: 770–82.
References for this review were identified from searches of 3 Nilsson P, Larsson E-M, Maly-Sundgren P, Perfekt
the extensive files of the authors; presentations at the R, Sandberg-Wollheim M. Predicting the outcome of optic
MAGNIMS Workshop, held in London in January 2004 (see neuritis—evaluation of risk factors after 30 years of follow up.
J Neurol (in press).
acknowledgments); access to PubMed citations between 4 Comi G, Filippi M, Barkhof F, et al. Effect of early interferon
1985 and December 2004; and references from relevant treatment on conversion to definite multiple sclerosis. Lancet 2001;
articles. A few abstracts and reports from the MAGNIMS 357: 1576–82.
5 Eriksson M, Andersen O, Runmarker B. Long-term follow-up of
meeting were also included. The final reference list was patients with clinically isolated syndromes, relapsing remitting and
generated from papers that were original and relevant to the secondary progressive multiple sclerosis. Mult Scler 2003; 9: 260–74.
topics covered in the review. 6 Weinshenker BG, Bass B, Rice GP, et al. The natural history of
multiple sclerosis: a geographically based study—2. Predictive
value of the early clinical course. Brain 1989; 112: 1419–28.
7 Kurtzke JF, Beebe GW, Nagler B, Kurland LT, Auth TL. Studies on
These data indicate that in a mixed CIS population, the the natural history of multiple sclerosis–8. Early prognostic features
of the later course of the illness. J Chronic Dis 1977; 30: 819–30.
number of baseline lesions predicts not only conversion
8 Miller DH, Hornabrook RW, Purdie G. The natural history of
to clinically definite MS, but also the likelihood of multiple sclerosis: a regional study with some longitudinal data.
developing mild and even moderate disability; note that J Neurol Neurosurg Psychiatry 1992; 55: 341–46.
all three deaths due to MS occurred in patients with four 9 Jacobs LD, Kaba SE, Miller CM, Priore RL, Brownscheidle CM.
Correlation of clinical, magnetic resonance imaging and
or more lesions at baseline.10 In the Amsterdam group, cerebrospinal fluid findings in optic neuritis. Ann Neurol 1997;
EDSS score 3 was reached in 48% of patients with four 41: 392–98.
or more lesions, and in 55% of patients with ten or more 10 Brex PA, Ciccarelli O, O’Riordan JI, et al. A longitudinal study of
abnormalities on MRI and disability from multiple sclerosis.
lesions.11 These data also suggest that the distribution of N Engl J Med 2002; 346: 158–64.
lesions may be important. Logistic regression analysis 11 Optic Neuritis Study Group. High- and low-risk profiles for the
has shown that the likelihood of obtaining EDSS 3 was development of multiple sclerosis within 10 years after optic
neuritis. Arch Ophthalmol 2003; 121: 944–49.
best predicted by the presence of two or more
12 Minneboo A, Barkhof F, Polman CH, et al. Infratentorial lesions
infratentorial lesions.11 predict long term disability in patients with initial findings
suggestive of multiple sclerosis. Arch Neurol 2004; 61: 217–21.
Acknowledgments 13 Frederiksen JL, Larsson HB, Olesen J, Stigsby B. MRI, VEP, SEP
Part of the content of this review was drawn from presentations made at and biothesiometry suggest monosymptomatic acute optic neuritis
an International Workshop on Clinically Isolated Syndromes, held in to be a first manifestation of multiple sclerosis. Acta Neurol Scand
London, UK, in January 2004. The workshop was coordinated by 1991; 83: 343–50.
MAGNIMS, a European network of investigators who are interested in 14 Morrissey SP, Miller DH, Kendall BE, et al. The significance of
the application of MRI methods to MS. Sponsorship support was brain magnetic resonance imaging abnormalities at presentation
received from the MS Society of Great Britain and Northern Ireland, the with clinically isolated syndromes suggestive of multiple sclerosis.
Italian MS Society, and the National MS Society (US). Workshop Brain 1993; 116: 135–46.
participants were: O Andersen (Gothenburg), D Arnold (Montreal), 15 Soderstrom M, Lindqvist M, Hillert J, Kall TB, Link H. Optic
F Barkhof (Amsterdam), T Berger (Innsbruck), L Bo (Amsterdam), neuritis: findings on MRI, CSF examination and HLA class II
P Brex (London), A Compston (Cambridge), C Confavreux (Lyons), typing in 60 patients and results of a short-term follow-up.
Neurology 1994; 241: 391–97.
C Dalton (London), M Daumer (Munich), F Fazekas (Graz), K Fernando
(London), M Filippi (Milan), J Frederiksen (Copenhagen), S Hickman 16 Beck RW, Cleary PA, Trobe JD, et al. The effect of corticosteroids
for acute optic neuritis on the subsequent development of multiple
(London), R Hohlfeld (Munich), M Inglese (New York), M Johnson
sclerosis. N Engl J Med 1993; 329: 1764–69.
(Leeds), L Kappos (Basel), M Kupersmith (New York), C Lucchinetti
17 Jacobs L, Beck R, Simon J, et al. Intramuscular interferon beta-1a
(Rochester, Minnesota), G Lycklama (Amsterdam), H McFarland therapy initiated during the first demyelinating event in multiple
(Washington DC), P Matthews (Oxford), D Miller (London), sclerosis. N Engl J Med 2000; 343: 898–904.
X Montalban (Barcelona), G Plant (London), C Polman (Amsterdam), S 18 Tintorè M, Rovira A, Rio J. Is optic neuritis more benign than other
Reingold (New York), H Roed (Copenhagen), M Rocca (Milan), M Ron first attacks in multiple sclerosis? Ann Neurol 2005; 57: 210–15.
(London), A Rovira (Barcelona), M Rovaris (Milan), M Sandberg- 19 O’Riordan JI, Thompson AJ, Kingsley DPE, et al. The
Wollheim (Lund), J Simon (Denver), F Sellebjerg (Copenhagen), A prognostic value of brain MRI in clinically isolated syndromes of
Thompson (London), M Tintoré (Barcelona), A Toosy (London), B the central nervous system: a 10 year follow-up. Brain 1998;
Weinshenker (Rochester, Minnesota), and J Wolinsky (Houston). 121: 495–503.
20 Callanan MM, Logsdail SJ, Ron MA, Warrington EK. Cognitive
Authors’ contributions impairment in patients with clinically isolated lesions of the type
We all contributed equally to the writing of this review. seen in multiple sclerosis: a psychometric and MRI study.
Brain 1989; 112: 361–74.
Conflicts of interest 21 Feinstein A, Kartsounis LD, Miller DH, Youl BD, Ron MA.
We have no conflicts of interest. Clinically isolated lesions of the type seen in multiple sclerosis:
a cognitive, psychiatric, and MRI follow up study.
Role of the funding source J Neurol Neurosurg Psychiatry 1992; 55: 869–76.
No funding source was involved in the preparation of this review or in 22 Filippi M, Tortorella C, Rovaris M, et al. Changes in the normal
the decision to submit it for publication. appearing brain tissue and cognitive impairment in multiple
sclerosis. J Neurol Neurosurg Psychiatry 2000; 68: 157–61.
References 23 Zivadinov R, Sepcic J, Nasuelli D, et al. A longitudinal study of
1 Confavreux C, Vukusic S, Moreau T, Adeleine P. Relapses and brain atrophy and cognitive disturbances in the early phase of
progression of disability in multiple sclerosis. N Engl J Med 2000; relapsing-remitting multiple sclerosis. J Neurol Neurosurg Psychiatry
343: 1430–38. 2001; 70: 773–80.

http://neurology.thelancet.com Vol 4 May 2005 287

 Review

24 Gadea M, Martínez-Bisbal MC, Marti-Bonmatí L, et al. 48 Lee KH, Hashimoto SA, Hooge JP, et al. Magnetic resonance
Spectroscopic axonal damage of the right locus coeruleus relates to imaging of the head in the diagnosis of multiple sclerosis: a
selective attention impairment in early stage relapsing-remitting prospective 2-year follow-up with comparison of clinical evaluation,
multiple sclerosis. Brain 2004; 127: 89–98. evoked potentials, oligoclonal banding, and CT. Neurology 1991;
25 Ranjeva JP, Audoin B, Duong MVA, et al. MR study of the 41: 657–60.
influence of tissue damage and cortical reorganization on PASAT 49 Fazekas F, Offenbacher H, Fuchs S, et al. Criteria for an increased
performance at the earliest stage of multiple sclerosis. Neurology specificity of MRI interpretation in elderly subjects with suspected
2004; 62 (suppl 5): A92–93. multiple sclerosis. Neurology 1988; 38: 1822–25.
26 Lucchinetti C, Bruck W, Parisi J, Scheithauer B, Rodriguez M, 50 Offenbacher H, Fazekas F, Schmidt R, et al. Assessment of MRI
Lassmann H. Heterogeneity of multiple sclerosis lesions: criteria for a diagnosis of MS. Neurology 1993; 43: 905–09.
Implications for the pathogenesis of demyelination. Ann Neurol 51 Tas MW, Barkhof F, van Walderveen MA, Polman CH,
2000; 47: 707–17. Hommes OR, Valk J. The effect of gadolinium on the sensitivity
27 Barnett MH, Prineas JW. Relapsing-remitting multiple sclerosis: and specificity of MR in the initial diagnosis of multiple sclerosis.
pathology of the newly forming lesion. Ann Neurol 2004; Am J Neuroradiol 1995; 2: 259–64.
55: 458–68. 52 Tintoré M, Rovira A, Martínez MJ, et al. Isolated demyelinating
28 Trapp BD. Pathogenesis of multiple sclerosis: the eyes only see syndromes: comparison of different MRI criteria to predict
what the mind is prepared to comprehend. Ann Neurol 2004; conversion to clinically definite multiple sclerosis. Am J Neuroradiol
55: 455–57. 2000; 21: 702–06.
29 Lumsden CE. The neuropathology of multiple sclerosis. 53 McDonald WI, Compston A, Edan G, et al. Recommended
In: Vinken PJ, Bruyn GW, eds. Handbook of clinical neurology, diagnostic criteria for multiple sclerosis: guidelines from the
volume 9. Amsterdam: Elsevier, 1970: 217–309. International Panel on the Diagnosis of Multiple Sclerosis.
Ann Neurol 2001; 50: 121–27.
30 Kidd D, Barkhof F, McConnell R, Algra PR, Allen IV, Revesz T.
Cortical lesions in multiple sclerosis. Brain 1999; 122: 17–26. 54 Tintoré M, Rovira A, Rio J, et al. New diagnostic criteria for
multiple sclerosis: application in first demyelinating episode.
31 Peterson JW, Bo L, Mork S, Chang A, Trapp BD. Transected
Neurology 2003; 60: 27–30.
neuritis, apoptotic neurons, and reduced inflammation in cortical
multiple sclerosis lesions. Ann Neurol 2001; 50: 389–400. 55 Dalton CM, Brex PA, Miszkiel KA, et al. Application of the new
McDonald criteria to patients with clinically isolated syndromes
32 Sandberg-Wollheim M, Bynke H, et al. A long-term prospective
suggestive of multiple sclerosis. Ann Neurol 2002; 52: 47–53.
study of optic neuritis: evaluation of risk factors. Ann Neurol 1990;
27: 386–93. 56 Dalton CM, Brex PA, Miszkiel KA, et al. New T2 lesions to enable
an earlier diagnosis of multiple sclerosis in clinically isolated
33 Sellebjerg F, Jensen CV, Christiansen M. Intrathecal IgG synthesis
syndromes. Ann Neurol 2003; 53: 673–76.
and autoantibody secreting cells in multiple sclerosis.
J Neuroimmunol 2000; 108: 207–15. 57 Barkhof F, Rocca M, Francis G, et al. Validation of diagnostic
magnetic resonance imaging criteria for multiple sclerosis and
34 Berger T, Rubner P, Schautzer F, et al. Antimyelin antibodies as a response to interferon beta1a. Ann Neurol 2003; 53: 718–24.
predictor of clinically definite multiple sclerosis after a first
demyelinating event. N Eng J Med 2003; 349: 139–45. 58 Miller DH, Filippi M, Fazekas F, et al. The role of MRI within
diagnostic criteria for multiple sclerosis: a critique. Ann Neurol
35 Gaertner S, de Graaf KL, Greve B, Weissert R. Antibodies against 2004; 56: 273–78.
glycosylated native MOG are elevated in patients with multiple
59 Sastre-Garriga J, Tintoré M, Rovira A, et al. Specificity of Barkhof
sclerosis. Neurology 2004; 63: 2381–83.
criteria in predicting conversion to multiple sclerosis when applied
36 Lampasona V, Franciotta D, Furlan R, et al. Similar low frequency to clinically isolated brainstem syndromes. Arch Neurol 2004;
of anti-MOG IgG and IgM in MS patients and healthy subjects. 61: 222–24.
Neurology 2004; 62: 2092–94.
60 Miller DH, Ormerod IEC, Rudge P, Kendall BE, Moseley IF,
37 Lim ET, Berger T, Reindl M, et al. Anti-myelin antibodies do not McDonald WI. The early risk of multiple sclerosis following
allow earlier diagnosis of multiple sclerosis. Mult Scler (in press). isolated acute syndromes of the brainstem and spinal cord.
38 Jensen J, Langkilde AR, Fenst C, et al. CD4 T cell activation and Ann Neurol 1989; 26: 635–39.
disease activity at onset of multiple sclerosis. J Neuroimmunol 2004; 61 Ford B, Tampieri D, Francis G. Long-term follow-up of acute
149: 202–09. partial transverse myelopathy. Neurology 1992; 42: 250–52.
39 Sorensen TL, Roed H, Sellebjerg F. Monosymptomatic optic 62 Tartaglino LM, Friedman DP, Flanders AE, Lublin FD, Knobler RL,
neuritis: chemokine receptor CXCR3 and its ligands. Liem M. Multiple sclerosis in the spinal cord: MR appearance and
Br J Ophthalmol 2004; 88: 1146–48. correlation with clinical parameters. Radiology 1995; 195: 725–32.
40 Bielekova B, Martin R. Development of biomarkers in multiple 63 O’Riordan JI, Losseff NA, Phatouros C, et al. Asymptomatic spinal
sclerosis. Brain 2004; 127: 1463–78. cord lesions in clinically isolated optic nerve, brain stem and spinal
41 Hickman SJ, Dalton CM, Miller DH, Plant GT. Management of cord syndromes of demyelination. J Neurol Neurosurg Psychiatry
acute optic neuritis. Lancet 2002; 360: 1953–62. 1998; 64: 353–57.
42 Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for 64 Bot JC, Barkhof F, Polman CH, et al. Spinal cord abnormalities in
multiple sclerosis: guidelines for research proposals. Ann Neurol recently diagnosed MS patients: added value of spinal MRI
1983; 13: 227–31. examinations. Neurology 2004; 62: 226–33.
43 Jacobs L, Kinkel PR, Kinkel WR. Silent brain lesions in patients 65 Dalton CM, Brex PS, Miszkiel KM, et al. Spinal cord MRI in
with isolated idiopathic optic neuritis: a clinical and nuclear clinically isolated optic neuritis. J Neurol Neurosurg Psychiatry 2003;
magnetic resonance imaging study. Arch Neurol 1986; 43: 452–55. 74: 1577–80.
44 Ormerod IE, Bronstein A, Rudge P, et al. Magnetic resonance 66 Tintore M, Rovira A, Rio J, et al. The prognostic value of brain
imaging in clinically isolated lesions of the brain stem. magnetic resonance imaging and oligoclonal bands in clinically
J Neurol Neurosurg Psychiatry 1986; 49: 737–43. isolated syndromes: 5 year follow up. Mult Scler 2004; 10 (suppl 2):
45 Miller DH, McDonald WI, Blumhardt LD, et al. Magnetic S187.
resonance imaging in isolated non-compressive spinal cord 67 Miller DH, Ormerod IEC, McDonald WI, et al. The early risk of
syndromes. Ann Neurol 1987; 22: 714–23. multiple sclerosis after optic neuritis. J Neurol Neurosurg Psychiatry
46 Barkhof F, Filippi M, Miller DH, et al. Comparison of MRI criteria 1988; 51: 1569–71.
at first presentation to predict conversion to clinically definite 68 Frohman E, Goodin DS, Calabresi PA, et al. The utility of MRI in
multiple sclerosis. Brain 1997; 120: 2059–69. suspected MS: report of the Therapeutics and Technology
47 Paty DW, Oger JJ, Kastrukoff LF, et al. MRI in the diagnosis of MS: Assessment Subcommittee of the American Academy of
a prospective study with comparison of clinical evaluation, evoked Neurology. Neurology 2003; 61: 602–11.
potentials, oligoclonal banding, and CT. Neurology 1988; 69 Optic Neuritis Study Group. Neurologic Impairment 10 years after
38: 180–85. optic neuritis. Arch Neurol 2004; 61: 1386–89.

288 http://neurology.thelancet.com Vol 4 May 2005