doi:10.

1093/brain/awy006 BRAIN 2018: 141; 1075–1084 | 1075

The value of oligoclonal bands in the multiple

sclerosis diagnostic criteria
Georgina Arrambide,1,* Mar Tintore,1,* Carmen Espejo,1 Cristina Auger,2 Mireia Castillo,1
Jordi Rı́o,1 Joaquı́n Castilló,1 Angela Vidal-Jordana,1 Ingrid Galán,1 Carlos Nos,1
Raquel Mitjana,2 Patricia Mulero,1 Andrea de Barros,2 Breogán Rodrı́guez-Acevedo,1

Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024

Luciana Midaglia,1 Jaume Sastre-Garriga,1 Alex Rovira,2 Manuel Comabella1 and
Xavier Montalban1,3

*These authors contributed equally to this work.

The presence of oligoclonal bands in clinically isolated syndromes is an independent risk factor for developing multiple sclerosis
and has been largely excluded from the more recent multiple sclerosis diagnostic criteria. Therefore, our objective was to explore
the value of oligoclonal bands in the context of the 2010 McDonald criteria, especially in patients fulfilling exclusively dissem-
ination in space at baseline. For this purpose, we selected 566 patients from a clinically isolated syndrome inception cohort who
had IgG oligoclonal bands determination and sufficient data on baseline brain MRI to assess dissemination in space and time. We
excluded the cases already fulfilling both dissemination in space and time and divided the remaining 398 into ‘no dissemination in
space and time’ (n = 218), ‘dissemination in space’ (n = 164) and ‘dissemination in time’ (n = 16). We assessed Cox proportional
hazards regression models with 2010 McDonald as the outcome, using ‘no dissemination in space and time’ with 0 lesions and
negative oligoclonal bands as the reference for different subgroups according to oligoclonal bands status (positive/negative). To
assess the diagnostic properties, we selected cases with a follow-up 53 years or fulfilling 2010 McDonald within 3 years of the
clinically isolated syndrome (n = 314), and compared the performance of all ‘dissemination in space’ cases (n = 137) versus patients
with ‘dissemination in space’ and positive oligoclonal bands (n = 101). The remaining patients classified as fulfilling ‘dissemination
in time’ or ‘no dissemination in space and time’ were taken into account to calculate the diagnostic properties. The respective
adjusted hazard ratios (95% confidence interval) were 1.5 (0.4–5.7) for ‘no dissemination in space and time’ with 0 lesions and
positive oligoclonal bands, 3.1 (1.4–7.2) for ‘no dissemination in space and time’ with 51 lesions and negative oligoclonal bands,
7.4 (3.5–15.7) for ‘no dissemination in space and time’ with 51 lesions and positive oligoclonal bands, 10.4 (4.8–22.6) for
‘dissemination in space’ with negative oligoclonal bands, 15.3 (7.5–31.3) for ‘dissemination in space’ with positive oligoclonal
bands, and 9.1 (3.5–23.4) for ‘dissemination in time’ (not subdivided due to the sample size). The specificity for all cases with
‘dissemination in space’ was 80.6 and increased to 88.1 after selecting those with positive oligoclonal bands. According to these
results, we propose radiological dissemination in space at any time plus positive oligoclonal bands as an additional criterion for
diagnosing multiple sclerosis.

1 Servei de Neurologia-Neuroimmunologia, Centre d’Esclerosi Múltiple de Catalunya (Cemcat), Vall d’Hebron Institut de Recerca,
Hospital Universitari Vall d’Hebron, Universitat Autònoma de Barcelona, Barcelona, Spain
2 Section of Neuroradiology and Magnetic Resonance Unit, Department of Radiology (IDI), Vall d’Hebron Institut de Recerca,
Hospital Universitari Vall d’Hebron, Universitat Autònoma de Barcelona, Barcelona, Spain
3 Division of Neurology, University of Toronto, St. Michael’s Hospital, Toronto, Canada

Received September 5, 2017. Revised November 15, 2017. Accepted November 25, 2017. Advance Access publication February 16, 2018
ß The Author(s) (2018). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.
For Permissions, please email: journals.permissions@oup.com
1076 | BRAIN 2018: 141; 1075–1084
Correspondence to: Xavier Montalban
Centre d’Esclerosi Múltiple de Catalunya (Cemcat)
Edifici Cemcat
Hospital Universitari Vall d’Hebron
Pg. Vall d’Hebron 119-129
08035 Barcelona,
Spain
E-mail: xavier.montalban@cem-cat.org; MontalbanX@smh.ca
Keywords: multiple sclerosis; biomarkers; clinically isolated syndrome; oligoclonal bands; imaging
Abbreviations: CIS = clinically isolated syndrome; DIS = dissemination in space; DIT = dissemination in time
Introduction
The presence of IgG oligoclonal bands indicates an
increased risk for developing multiple sclerosis in patients
presenting with clinically isolated syndromes (CIS)
(McDonald et al., 2001; Polman et al., 2005; Tintore
et al., 2008, 2015; Zipoli et al., 2009) independently of
other risk factors (Tintore et al., 2008, 2015; Zipoli et al.,
2009). In fact, the 2001 McDonald criteria for the diag-
nosis of multiple sclerosis and their 2005 revision allowed
an alternative dissemination in space (DIS) criterion com-
prised by the presence of two or more T2 lesions on MRI
plus positive oligoclonal bands (McDonald et al., 2001;
Polman et al., 2005). However, in the 2010 revisions, it
was deemed inappropriate to further liberalize MRI re-
quirements in CSF-positive patients as the contribution
of oligoclonal bands status was not evaluated (Polman
et al., 2011). In this sense, a recent multicentre study
showed that the diagnostic performance of 2010
DIS plus positive oligoclonal bands was similar to DIS
alone (Huss et al., 2016), whereas another study in chil-
dren with optic neuritis demonstrated that the combin-
ation of MRI lesions and oligoclonal bands positivity
was indicative of having multiple sclerosis (Heussinger
et al., 2015).
Therefore, the objective of this study was to explore the
value of oligoclonal bands for multiple sclerosis diagnosis
in the context of the 2010 McDonald criteria, especially in
patients with a CIS typical for demyelination fulfilling only
DIS at baseline.
Materials and methods
Study cohort
This study was based on the Barcelona open CIS cohort at the
Multiple Sclerosis Centre of Catalonia, Vall d’Hebron
University Hospital, Barcelona.
This study received approval from the Clinical Research
Ethics Committee at Vall d’Hebron University Hospital
[EPA(AG)57/2013(3834)]. All patients signed a written in-
formed consent.
G. Arrambide et al.
Patients
CIS patients 550 years of age first seen within 3 months of
Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024
disease onset were included (Tintore et al., 2015).
Demographic data, CIS topography, and disability according
to the Expanded Disability Status Scale (EDSS) were recorded
at baseline. Follow-up was performed every 3–6 months, as-
sessing for relapses and disability according to the EDSS mea-
sured during stability periods.
IgG oligoclonal bands
Oligoclonal bands were determined in CSF and serum within
the first 3 months by agarose isoelectric focusing combined
with immunoblotting (Andersson et al., 1994; Freedman
et al., 2005).
MRI acquisition and analysis
The MRI protocol has been described elsewhere (Tintore et al.,
2015; Arrambide et al., 2017). Briefly, the baseline brain MRIs
were performed within 3–5 months of disease onset and
follow-up scans at 1 year and every 5 years thereafter.
Baseline spinal cord MRIs were initially done if the presenting
symptoms suggested a myelitis, and from 2007 they were ob-
tained in all CIS cases. Scans were obtained at 3.0 T since
2010 and at 1.5 T previously. Brain sequences included trans-
verse dual echo T2-weighted fast spin-echo, transverse and sa-
gittal T2-FLAIR, and transverse T1-weighted spin-echo. From
2001, the transverse T1-weighted sequence was systematically
repeated after gadolinium (Gd) injection (0.2 mmol/kg) in pa-
tients with lesions on T2-weighted images. Previous to that
year, Gd was seldom administered. Spinal cord sequences
included sagittal dual echo proton density/T2-weighted fast
spin-echo, sagittal short-tau inversion-recovery (STIR) and, in
patients with brain Gd T1-weighted sequences or spinal cord le-
sions, a Gd-enhanced sagittal T1-weighted. Axial T2-weighted
sequences were performed, covering segments showing
abnormalities on the sagittal images or with suspected involve-
ment based on clinical findings.
All sequences were acquired with a contiguous 3-mm section
thickness. MRI scans were assessed by one of two neuroradiol-
ogists blinded to clinical follow-up. In doubtful cases, the final
analysis was based on their consensus opinion. MRIs were
considered abnormal if one or more lesions were observed.
The number and location of lesions on T2-weighted images,
number of Gd-enhancing lesions, and number of active T2 le-
sions on brain MRI were scored. Lesion number (0, 1, 2–3,
43) and presence of Gd enhancement on spinal cord MRI
were scored.
Oligoclonal bands and McDonald criteria
Experimental design
From January 1995 to January 2016, we included 1214 pa-
tients in the CIS cohort. Of these, we excluded 71 (5.8%)
during follow-up for having presented a previous demyelinat-
ing attack (n = 13), exceeding the age limit (n = 4), exceeding
the entry window (n = 26), or reaching an alternative diagnosis
(n = 28). Of the remaining 1143 patients, we selected the study
groups as described below (Fig. 1).
Risk assessment group
From this cohort, we selected patients who had both oligoclonal
bands determination and sufficient information on the baseline
brain MRI to assess DIS and dissemination in time (DIT) ac-
cording to a modified version of the 2010 McDonald criteria
(Polman et al., 2011), taking into account the symptomatic
lesions (Brownlee et al., 2016; Tintore et al., 2016) (n = 566).
Of these patients, we excluded 168 (29.7%) already fulfilling
DIS plus DIT and divided the remaining 398 into cases fulfilling
‘DIS’ (n = 164), ‘DIT’ (n = 16), and ‘no DIS no DIT’ (n = 218).
Of the 577 non-selected patients out of 1143, 347 (60.1%) had
oligoclonal bands determination but none of them had Gd data
collected, mostly due to changes in the MRI protocol over time.
Diagnostic properties group
From the risk assessment cohort, we further selected cases with
a minimum follow-up of 3 years or fulfilling the 2010
CIS cohort
January 1995 - January 2016
n = 1214
n = 1143
n = 566
n = 398
No DIS no DIT: n = 218 DIS: n = 164
All DIS (n = 137) versus DIS with +OB (n =
101)
Figure 1 Algorithm showing the selection of patients from the CIS cohort. OB = oligoclonal bands; MS = multiple sclerosis.
BRAIN 2018: 141; 1075–1084 | 1077
McDonald multiple sclerosis criteria within 3 years of the
CIS (Swanton et al., 2007). Of the resultant 314 cases, ‘DIS’
was fulfilled in 137, ‘DIT’ in 12, and ‘no DIS no DIT’ in 165.
Statistical analysis
Descriptive statistics were performed on demographic and clin-
ical variables in both studied groups.
Risk assessment group
We performed uni- and multivariable Cox proportional haz-
ards regression analyses, all with 95% confidence intervals
(CI), with the modified 2010 McDonald criteria (fulfilled
either clinically or radiologically) as the outcome after dividing
Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024
the ‘no DIS no DIT’ cases into four subgroups: 0 lesions and
negative oligoclonal bands (n = 91), which was used as the
reference category; 0 lesions and positive oligoclonal bands
(n = 16); 51 lesions and negative oligoclonal bands (n = 60);
and 51 lesions and positive oligoclonal bands (n = 51). The
‘DIS’ subgroup was also divided according to oligoclonal
bands status (negative, n = 42; and positive, n = 122). Due to
the small number of cases fulfilling ‘DIT’ we did not divide this
subgroup any further.
Covariates included age, gender, CIS topography, and dis-
ease-modifying treatment as a binary, time-dependent variable.
Excluded: n = 71 (5.8%)
-Previous attack: n = 13
-Age >50 years: n = 4
-Exceeded entry window: n = 26
-Other diagnoses: n = 28
Selection criteria:
-Not missing OB data
-Baseline brain MRI with sufficient data
to establish DIS and DIT status
Excluding patients fulfilling 2010
McDonald MS: n = 168
Risk assessment
DIT: n = 16
-Follow-up ≥3 years
or
-2010 McDonald MS within 3 years of the CIS
Diagnostic properties
1078 | BRAIN 2018: 141; 1075–1084 G. Arrambide et al.

Diagnostic properties group
We assessed the diagnostic properties of all DIS cases, regard-
less of the oligoclonal bands status (n = 137), and compared
them to patients fulfilling DIS and with positive oligoclonal
bands (n = 101) by use of 2010 McDonald at 3 years as the
outcome. The remaining patients classified as fulfilling ‘DIT’ or
‘no DIS no DIT’ were taken into account when calculating the
diagnostic properties. The number of true positives (patients
fulfilling the different MRI-oligoclonal bands criteria and de-
veloping McDonald multiple sclerosis), false positives (patients
fulfilling the criteria but not developing McDonald multiple
sclerosis), true negatives (patients not fulfilling the criteria
and not developing McDonald multiple sclerosis), and false
negatives (patients not fulfilling the criteria but developing
McDonald multiple sclerosis) was determined. We then calcu-
similar characteristics, except for a longer follow-up, more
treated patients, and a higher proportion of patients de-
veloping multiple sclerosis.
The data according to the different DIS or DIT sub-
groups are detailed in Table 2. In both the risk assessment
and performance groups, the proportion of patients with
positive oligoclonal bands or starting a disease-modifying
treatment was higher in ‘DIS’ than in ‘DIT’ and ‘no DIS no
DIT’. Additionally, more ‘DIS’ and ‘DIT’ patients reached
the diagnosis of McDonald multiple sclerosis than ‘no DIS
no DIT’ cases, with ‘DIS’ patients doing so earlier than
‘DIT’ cases, although without statistical significance
[median (percentiles 25–75) 12.3 (7.4–20.5) versus 35.7
(16.8–47.6) months, respectively, P = 0.123].

Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024

lated the sensitivity [true positives/(true positives + false nega-
tives)], specificity [true negatives/(true negatives + false
positives)], accuracy [(true positives + true negatives)/total
tests], positive predictive value (PPV) [true positives/(true posi-
tives + false positives)], and negative predictive value (NPV) [true
negatives/(false negatives + true negatives)], all with 95% CIs.
Statistical tests were performed on the 0.05 level of signifi-
cance using the IBM SPSS Statistics (SPSS Inc., Chicago, IL,
USA), version 22.0.
Results
General characteristics
Table 1 shows the general characteristics of both the risk
assessment and the diagnostic performance groups. Relative
to the risk assessment group, the performance group had
Figure 2 shows the proportion of patients reaching the
outcome according to the MRI-oligoclonal bands sub-
groups, which was 460.0% for both ‘DIS’ subgroups
and for ‘no DIS no DIT’ with 51 lesions and positive
oligoclonal bands. Figure 3 shows the corresponding
median time to McDonald multiple sclerosis, which was
shorter for both ‘DIS’ subgroups (P = 0.284 for all). Of
note, the number of patients initiating DMT at any time
was 0/91 (0%) in ‘no DIS no DIT’ with 0 lesions and
negative oligoclonal bands, 1/16 (6.3%) in ‘no DIS no
DIT’ with 0 lesions and positive oligoclonal bands, 10/60
(16.7%) in ‘no DIS no DIT’ with 51 lesions and negative
oligoclonal bands, 24/51 (47.1%) in ‘no DIS no DIT’ with
51 lesions and positive oligoclonal bands, 22/42 (52.4%)
in ‘DIS and negative oligoclonal bands’, 81/122 (66.4%) in
‘DIS and positive oligoclonal bands’, and 5/16 (31.3%) in
‘DIT’ (P 5 0.0001).

Table 1 General demographic and clinical characteristics of the two studied patient groups

Risk assessment group Performance group

n = 398 n = 314
Females: n (%) 260 (65.3) 205 (65.3)
Mean (SD) age in years 32.6 (8.0) 32.6 (7.8)
CIS topography: n (%)
Optic nerve 146 (36.7) 108 (34.4)
Infratentorial 100 (25.1) 79 (25.2)
Spinal cord 103 (25.9) 86 (27.4)
Other 49 (12.3) 41 (13.1)
+ OB: n (%) 196 (49.2) 164 (52.2)
Abnormal brain MRI: n (%) 266 (66.8) 220 (70.1)
Abnormal spinal cord MRI: n (%) 74/195 (37.9) 57/136 (41.9)
+ Gd on brain MRI: n (%) 11 (2.8) 9 (2.9)
+ Gd on spinal cord MRI: n (%) 5/195 (2.6) 3/136 (2.2)
Mean (SD) follow-up in years 6.5 (4.1) 7.9 (3.4)
DMT: n (%) 143 (35.9) 130 (41.4)
Outcome second attack: n (%) 144 (36.2) 144 (45.9)
Median (percentiles 25–75) time to second attack in months 25.0 (7.2–49.1) 25.0 (7.2–49.1)
Outcome 2010 McDonald MS: n (%)a 180 (45.2) 180 (57.3)
Median (percentiles 25–75) time to 2010 McDonald MS in months 12.5 (9.8–33.5) 12.5 (9.8–33.5)

All MRI data were obtained assessing only baseline scans.

a
Modified version including the symptomatic lesions. The outcome can be either clinical or radiological.
DMT = disease-modifying treatment; MS = multiple sclerosis; OB = oligoclonal bands; SD = standard deviation.
Oligoclonal bands and McDonald criteria BRAIN 2018: 141; 1075–1084 | 1079

Table 2 General characteristics of the studied groups according to the fulfilment of the 2010 McDonald criteria at
baseline

Risk assessment group Diagnostic performance group

n = 398 n = 314
No DIS DIS DIT No DIS DIS DIT
no DIT no DIT
n = 218 n = 164 n = 16 n = 165 n = 137 n = 12
a
Females: n (%) 133 (61.0) 116 (70.7) 11 (68.8) 103 (62.4) 95 (69.3) 7 (58.3)
Mean (SD) age in yearsb 31.6 (7.9) 33.9 (8.1) 33.1 (5.8) 31.9 (7.6) 33.5 (8.1) 32.9 (6.4)
CIS topography: n (%)c
Optic nerve 88 (40.4) 55 (33.5) 3 (18.8) 63 (38.2) 43 (31.4) 2 (16.7)
Infratentorial 51 (23.4) 45 (27.4) 4 (25.0) 39 (23.6) 36 (26.3) 4 (33.3)
Spinal cord 54 (24.8) 43 (26.2) 6 (37.5) 44 (26.7) 38 (27.7) 4 (33.3)
Other 25 (11.5) 21 (12.8) 3 (18.8) 19 (11.5) 20 (14.6) 2 (16.7)

Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024

+ OB: n (%)d 67 (30.7) 122 (74.4) 7 (43.8) 59 (35.8) 101 (73.7) 4 (33.3)
Abnormal brain MRI: n (%)d 91 (41.7) 164 (100.0) 11 (68.8) 74 (44.8) 137 (100.0) 9 (75.0)
Abnormal spinal cord MRI: n (%)d,e 20/105 (19.0) 49/82 (59.8) 5/8 (62.5) 19/71 (26.8) 35/59 (59.3) 3/6 (50.0)
+ Gd on brain MRI: n (%)d 0 (0) 0 (0) 11 (68.8) 0 (0) 0 (0) 9 (75.0)
+ Gd on spinal cord MRI: n (%)d 0/105 (0) 0/82 (0) 5/8 (62.5) 0/71 (0) 0/59 (0) 3/6 (50.0)
Mean (SD) follow-up in yearsf 6.3 (3.9) 6.9 (4.3) 6.3 (4.2) 7.8 (3.2) 8.0 (3.8) 8.0 (3.3)
DMT: n (%)d 35 (16.1) 103 (62.8) 5 (31.3) 34 (20.6) 91 (66.4) 5 (41.7)
Outcome second attack: n (%)d 53 (24.3) 83 (50.6) 8 (50.0) 53 (32.1) 83 (60.6) 8 (66.7)
Median (percentiles 25–75) time to 34.5 (7.5–55.5) 21.0 (7.5–41.4) 35.7 (16.8–47.6) 34.5 (7.5–55.5) 21.0 (7.5–41.4) 35.7 (16.8–47.6)
second attack in monthsg
Outcome 2010 McDonald MS: n (%)d,h 60 (27.5) 111 (67.7) 9 (56.3) 60 (36.4) 111 (81.0) 9 (75.0)
Median (percentiles 25–75) time to 2010 24.5 (7.4–55.5) 12.3 (7.4–20.5) 35.7 (16.8–47.6) 24.5 (7.4–55.5) 12.3 (7.4–20.5) 35.7 (16.8–47.6)
McDonald MS in monthsi

All MRI data were obtained assessing only baseline scans.

a
Risk assessment group, P = 0.129; diagnostic performance group, P = 0.411.
b
Risk assessment group, P = 0.021; diagnostic performance group, P = 0.281.
c
Risk assessment group, P = 0.501; diagnostic performance group, P = 0.748.
d
P 5 0.0001. In the case of the median time to 2010 McDonald MS, the statistical difference is for DIS in relation to the no DIS no DIT and DIT groups.
e
Diagnostic performance group, P = 0.001.
f
Risk assessment group, P = 0.352; diagnostic performance group, P = 0.836.
g
Risk assessment and diagnostic performance groups, P = 0.112.
h
Modified version including the symptomatic lesions. The outcome can be either clinical or radiological.
i
P = 0.123.
DMT = disease-modifying treatment; MS = multiple sclerosis; OB = oligoclonal bands.

Risk assessment group Discussion

Figure 4 shows the uni- and multivariable results for the
different studied subgroups. We observed a stepwise in-
crease in the risk of developing multiple sclerosis over
time that was not only dependent on DIS fulfilment or on
the presence of T2 lesions, but also on oligoclonal bands
status. For instance, the risk was already high for patients
with ‘DIS and negative oligoclonal bands’, but increased
further in ‘DIS with positive oligoclonal bands’. If consider-
ing all ‘DIS’ patients regardless of oligoclonal bands status,
the adjusted hazard ratio was 9.5 (95% CI 5.0–18.1).
Diagnostic properties group
Table 3 shows the diagnostic properties at 3 years for all
‘DIS’ and for ‘DIS with positive oligoclonal bands’. The
specificity for McDonald multiple sclerosis was 80.6 for
all ‘DIS’ cases, and increased to 88.1 when selecting only
the patients with positive oligoclonal bands.
The 2010 McDonald criteria are highly specific for multiple
sclerosis (Swanton et al., 2007; Rovira et al., 2009;
Montalban et al., 2010). Nevertheless, it is not uncommon
to identify patients presenting a typical CIS and demyelinat-
ing lesions who do not fulfil these criteria at baseline
(Caucheteux et al., 2015; Huss et al., 2016). This, in
part, could be due to the timing of the baseline MRI,
which could modify the possibilities of demonstrating
DIT (Rovira et al., 2009). Therefore, identifying an alter-
native criterion to diagnose multiple sclerosis at baseline in
these cases, without compromising specificity, becomes rele-
vant. Additionally, the demonstration of CSF oligoclonal
bands, indicating chronic inflammatory activity within the
CNS, increases the diagnostic confidence. In the present
study, the median time to reach the diagnosis of multiple
sclerosis was approximately 1 year after the CIS for both
‘DIS’ subgroups, and the proportion of patients doing so
1080 | BRAIN 2018: 141; 1075–1084 G. Arrambide et al.

CIS
n = 398

No DIS no DIT DIS

DIT
n = 218 n = 164
n = 16 (4.0%)
(54.8%) (41.2%)

0 lesions and ≥1 lesions and ≥1 lesions and

0 lesions and Negative OB
negative OB negative OB positive OB Positive OB
positive OB n = 42
n = 91 n = 60 n = 51 n = 122 (74.4)
n = 16 (7.3%) (25.6%)
(41.3%) (27.5%) (23.4%)

Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024

McDonald MS McDonald MS McDonald MS McDonald MS
McDonald MS McDonald MS McDonald MS
n = 16 n = 32 n = 26 n = 85
n = 9 (9.9%) n = 3 (18.8%) n = 9 (56.3%)
(26.7%) (62.7%) (61.9%) (69.7%)

Figure 2 Proportion of patients reaching the 2010 McDonald multiple sclerosis diagnosis in the different studied subgroups. In
the DIT group, seven (43.8%) patients had positive oligoclonal bands (OB) and four of them (57.1%) reached a diagnosis of McDonald multiple
sclerosis, compared to 5/9 (55.6%) with negative oligoclonal bands. MS = multiple sclerosis.

175 44.0 60.2 18.9 18.9 13.1 12.2 29.2

(37.6-51.4) (33.6-94.9) (8.6-61.1) (7.3-57.4) (9.7-21.1) (11.1-14.4) (12.1-42.1)

150
Time to 2010 McDonald MS in months

125

100

75

50

25

No DIS no DIT with 0 No DIS no DIT with 0 No DIS no DIT with >=1 No DIS no DIT with >=1 DIS and -OB DIS and +OB DIT
lesions and -OB lesions and +OB lesions and -OB lesions and +OB

Figure 3 Median (percentiles 25–75) time to McDonald multiple sclerosis in months according to each studied subgroup. Of
note, during the first 6 months, 12 patients with DIS and positive oligoclonal bands (OB) fulfilled McDonald multiple sclerosis compared to five
patients with DIS with negative oligoclonal bands; during months 6–12, the outcome was fulfilled by 26 and 2, respectively; and during the 12–13
month period by 16 and 6, respectively. Notice that at the end of the 6–12 month and throughout the 12–13 month periods the number of
patients fulfilling the outcome increases greatly also due to DIT fulfilment on the 12-month follow-up MRI. MS = multiple sclerosis.

was 7.8–12.0% higher if oligoclonal bands were present, sclerosis was diagnosed in 59 more cases in the ‘DIS with
depending on the studied group (risk assessment and per- positive oligoclonal bands’ subgroup, of which 38 did so
formance groups, respectively). However, given the differ- during the first 12 months of disease evolution, compared
ence in the number of patients in each category, when to seven in the ‘DIS with negative oligoclonal bands’ sub-
looking at the absolute numbers, McDonald multiple group, suggesting that in patients already fulfilling DIS, the
Oligoclonal bands and McDonald criteria BRAIN 2018: 141; 1075–1084 | 1081

A n HR, 95% CI P

0 lesions, negative OB 91 1

0 lesions, positive OB 16 1.5 (0.4-5.6) 0.536

No DIS no DIT, >=1 lesions, negative OB

60 2.7 (1.2-6.2) 0.015

No DIS no DIT, >=1 lesions, positive OB

51 7.0 (3.3-14.7) <0.0001

DIS, negative OB 42 8.5 (4.0-18.3) <0.0001

DIS, positive OB 122 12.0 (6.0-23.9) <0.0001

Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024

DIT 16 7.8 (3.1-19.6) <0.0001

5 10 15 20 25

B n aHR (95% CI) P

0 lesions, negative OB 91 1

0 lesions, positive OB 16 1.5 (0.4-5.7) 0.517

No DIS no DIT, >=1 lesions, negative OB 60 3.1 (1.4-7.2) 0.007

No DIS no DIT, >=1 lesions, positive OB 51 7.4 (3.5-15.7) <0.0001

42 10.4 (4.8-22.6) <0.0001

DIS, negative OB

122 15.3 (7.5-31.3) <0.0001

DIS, positive OB

DIT 16 9.1 (3.5-23.4) <0.0001

10 20 30 40

Figure 4 Hazard ratios (A) and adjusted hazard ratios (B) for fulfilling 2010 McDonald multiple sclerosis over time. Adjusted by
sex, age, CIS topography, and disease-modifying treatment before McDonald multiple sclerosis. (a)HR = (adjusted)hazard ratio; OB = oligoclonal
bands.

Table 3 Diagnostic properties in patients fulfilling DIS

n = 314
n McDonald MS at 3 years n (%) Sensitivity Specificity Accuracy PPV NPV
(95% CI) (95% CI) (95% CI) (95% CI) (95% CI)
All DIS 137 111/137 61.7 80.6 69.7 81.0 61.0
(81.0) (54.1–68.8) (72.9–86.9) (64.3–74.8) (74.8–86.0) (56.1–65.7)
DIS with + OB 101 85/101 47.2 88.1 64.6 84.2 55.4
(84.2) (39.8–54.8) (81.3–93.0) (59.1–69.9) (76.6–89.6) (51.6–59.1)

Outcome: 2010 McDonald multiple sclerosis at 3 years.

MS = multiple sclerosis; PPV = positive predictive value; NPV = negative predictive value; OB = oligoclonal bands.

presence of oligoclonal bands could indicate a shorter time of fulfilling McDonald multiple sclerosis 4.9-fold over DIS
to McDonald multiple sclerosis. Furthermore, the multi- with negative oligoclonal bands. A recent study showed
variable regression analyses showed that combining DIS that 30.0% more patients with positive oligoclonal bands
with the presence of oligoclonal bands increased the risk fulfilled the 2010 McDonald criteria and did so earlier than
1082 | BRAIN 2018: 141; 1075–1084
patients with negative oligoclonal bands (Huss et al.,
2016). The proportion of patients with positive oligoclonal
bands having multiple sclerosis was even greater in a multi-
centre study of children with optic neuritis, especially when
combined with an abnormal brain MRI (i.e. one or more
demyelinating lesions) (Heussinger et al., 2015). As for the
risk of developing multiple sclerosis over time, Huss and
colleagues demonstrated a 2-fold increase in patients with
positive oligoclonal bands in their univariable analysis,
which is in-line with previous findings (Tintore et al.,
2008, 2015; Huss et al., 2016). In their multivariable ana-
lyses, Heussinger and colleagues (Heussinger et al., 2015)
demonstrated that in patients with both an abnormal MRI
and presence of oligoclonal bands at baseline, the risk for
developing multiple sclerosis was more than 20 times
higher than in patients with negative findings on MRI
and CSF. Although the results from these two studies do
demonstrate the value of oligoclonal bands in the general
context of the 2010 McDonald criteria, their results were
not split according to the subgroups hereby presented,
which could aid in discriminating those in which the pres-
ence of oligoclonal bands could be more valuable, like in
patients already fulfilling DIS. Besides, our results may not
be easily comparable to the findings by Heussinger et al.
(2015), given patients with alternative diagnoses were
included in their non-multiple sclerosis subgroup whereas
we excluded such cases from our database. Taken together,
all these results suggest that patients with CIS fulfilling DIS
at baseline will probably develop multiple sclerosis early in
the disease course and the risk of doing so increases if
oligoclonal bands are positive, even more so than DIT
cases in our study.
Concerning the diagnostic properties, the specificity for
multiple sclerosis diagnosis at 3 years was higher for DIS
cases with positive oligoclonal bands than for all DIS cases,
regardless of oligoclonal bands status. This is in contrast to
the results reported by Huss and colleagues (Huss et al.,
2016), which showed similar results for both. This differ-
ence could be due to the time-point established to assess
the outcome, which was 1 year longer in our study. A
high specificity is an important aspect of performance in
multiple sclerosis diagnosis considering the potential initi-
ation of DMT. Such high specificity was not compromised
when the current, simpler criteria for DIS were first as-
sessed, as long as DIT was also present. However, the
lower specificity of DIS alone (59.0%) suggested that
52 lesions in 52 typical locations were insufficient to
diagnose multiple sclerosis (Swanton et al., 2007). In
this sense, our study shows that adding positive oligoclo-
nal bands to DIS increases the specificity of multiple scler-
osis diagnosis, suggesting ‘DIS plus positive oligoclonal
bands’ could be used as an alternative to diagnose mul-
tiple sclerosis if DIT is not fulfilled at baseline.
Of note, the ‘no DIS no DIT’ subgroup with 51 lesions
and positive oligoclonal bands had a higher risk for fulfill-
ing McDonald multiple sclerosis than ‘no DIS no DIT’ with
51 lesions and negative oligoclonal bands, and had a very
G. Arrambide et al.
similar proportion of patients reaching the outcome as that
of ‘DIS with negative oligoclonal bands’. Previous studies
have shown that the proportion of patients developing a
second demyelinating attack was higher in patients not ful-
filling the Barkhof-Tintore criteria but presenting with posi-
tive oligoclonal bands compared to those with negative
oligoclonal bands (Tintore et al., 2008; Zipoli et al.,
2009), and similar results were reported for patients with
0 lesions and positive oligoclonal bands when assessing the
2010 criteria (Huss et al., 2016). Additionally, it has been
demonstrated that one single lesion, whether asymptomatic
or not, increases the risk of developing multiple sclerosis in
typical CIS (Brownlee et al., 2016; Tintore et al., 2016),
and that patients with a lower lesion load require a longer
Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024
time to develop multiple sclerosis (Fisniku et al., 2008;
Brownlee et al., 2016; Tintore et al., 2016). Considering
the possible added value of positive oligoclonal bands in
such instances, a previous study in CIS patients not fulfill-
ing the 2010 DIS criteria showed that combinations of
other predictive factors, including oligoclonal bands,
increased the risk of multiple sclerosis 2.7–3.6 times
(Ruet et al., 2014). Nevertheless, relying on very few le-
sions without DIT to diagnose multiple sclerosis could be
too liberal and lead to diagnostic errors in the daily clinical
practice. In such cases, considering the presence of oligo-
clonal bands could be relevant due to the increased risk for
developing multiple sclerosis independently of MRI findings
(Tintore et al., 2008, 2015; Dobson et al., 2013).
Therefore, future studies should focus on these subgroups
to further evaluate their specificity and/or risk of misdiag-
nosis, always taking into account that these criteria should
be applied to CIS suggestive of multiple sclerosis.
One limitation of this study is the sample size of the DIT
subgroup, which could influence both its results and the
possibility to assess the role of oligoclonal bands in these
cases. Another limitation is the lower number of spinal
cord MRIs performed, which could lead to the misclassifi-
cation of some cases among the different subgroups.
However, given the moderate added value of spinal cord
lesions to multiple sclerosis diagnosis at baseline in our
cohort (Arrambide et al., 2017), the number of potential
misclassifications is probably low. One more limitation is
the inclusion of patients starting treatment before reaching
the studied outcome. Whereas this is corrected by adding
DMT as a time-dependent variable in the multivariable
analysis for risk assessment, DMT initiation could have a
potential confounding effect in the analysis concerning the
diagnostic properties. Nevertheless, we considered that
excluding the patients treated before reaching the outcome
would result in a selection bias given their high proportion
in the DIS subgroups, and therefore decided to include
them all in the present study. A likely limitation, in general,
to the usefulness of oligoclonal bands in multiple sclerosis
diagnosis, is the determination technique itself, which is
rater-dependent (Freedman et al., 2005). However, a pos-
sible future alternative is the measurement of CSF kappa-
free light chains, which is a rapid, quantitative, and easy to
Oligoclonal bands and McDonald criteria
standardize tool, almost equal to oligoclonal bands with
regard to diagnostic sensitivity and specificity in patients
with early multiple sclerosis (Senel et al., 2014; Presslauer
et al., 2016; Voortman et al., 2017).
Importantly, using oligoclonal bands to diagnose multiple
sclerosis is not novel: the Poser criteria already proposed
that, in CIS patients, laboratory-supported definite multiple
sclerosis was fulfilled in those with either clinical evidence of
two lesions and positive CSF findings, or with clinical evi-
dence of one lesion, paraclinical evidence of another lesion,
and positive CSF findings (Poser et al., 1983). Besides, the
2001 McDonald criteria for the diagnosis of multiple scler-
osis and their 2005 revision allowed an alternative DIS cri-
terion comprised by the presence of two or more T2 lesions
on MRI plus positive oligoclonal bands (McDonald et al.,
2001; Polman et al., 2005). More recent studies and ours
present further evidence of the role of oligoclonal bands in
multiple sclerosis diagnosis, which is, to date, the second
clearest diagnostic marker after MRI. In conclusion, accord-
ing to these results, we propose MRI DIS at any time plus
positive oligoclonal bands as an additional criterion for mul-
tiple sclerosis diagnosis in patients presenting with a CIS
typical for demyelination.
Acknowledgements
The authors thank Susana Otero (Department of Epide-
miology and Servei de Neurologia-Neuroimmunologia,
Centre d’Esclerosi Múltiple de Catalunya, Vall d’Hebron
University Hospital, Barcelona, Spain) and Santiago
Pérez-Hoyos (Statistics and Bioinformatics Unit, Vall
d’Hebron Institut de Recerca, Barcelona, Spain) for statis-
tical analysis support.
Funding
This work was supported by the ‘Fondo de Investigación
Sanitaria’ (FIS) of the Ministry of Economy and
Competitiveness of Spain (grant PI14/01439 awarded to
X.M.), The ‘Red Española de Esclerosis Múltiple
(REEM)’ (RD12/0032; RD16/0015) sponsored by the
‘Fondo de Investigación Sanitaria’ (FIS) of the Ministry of
Economy and Competitiveness of Spain, and the ‘Ajuts per
donar Suport als Grups de Recerca de Catalunya (2014
SGR 1082)’ sponsored by the ‘Agència de Gestió d’Ajuts
Universitaris i de Recerca’ (AGAUR) of the Generalitat de
Catalunya in Spain.
References
Andersson M, Alvarez-Cermeno J, Bernardi G, Cogato I, Freman P,
Frederiksen J, et al. Cerebrospinal fluid in the diagnosis of multiple
sclerosis: a consensus report. J Neurol Neurosurg Psychiatry 1994;
57: 897–902.
BRAIN 2018: 141; 1075–1084 | 1083
Arrambide G, Rovira A, Sastre-Garriga J, Tur C, Castilló J, Rı́o J,
et al. Spinal cord lesions: a modest contributor to diagnosis in clin-
ically isolated syndromes but a relevant prognostic factor. Mult Scler
2017, in press. doi: 10.1177/1352458517697830.
Brownlee WJ, Swanton JK, Miszkiel KA, Miller DH, Ciccarelli O.
Should the symptomatic region be included in dissemination in
space in MRI criteria for MS? Neurology 2016; 87: 680–83.
Caucheteux N, Maarouf A, Genevray M, Leray E, Deschamps R,
Chaunu MP, et al. Criteria improving multiple sclerosis diagnosis
at the first MRI. J Neurol 2015; 262: 979–87.
Dobson R, Ramagopalan S, Davis A, Giovannoni G. Cerebrospinal
fluid oligoclonal bands in multiple sclerosis and clinically isolated
syndromes: a meta-analysis of prevalence, prognosis and effect of
latitude. J Neurol Neurosurg Psychiatry 2013; 84: 909–14.
Fisniku LK, Brex PA, Altmann DR, Miszkiel KA, Benton CE, Lanyon
R, et al. Disability and T2 MRI lesions: a 20-year follow-up of
patients with relapse onset of multiple sclerosis. Brain 2008; 131:
Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024
808–17.
Freedman MS, Thompson EJ, Deisenhammer F, Giovannoni G,
Grimsley G, Keir G, et al. Recommended standard of cerebrospinal
fluid analysis in the diagnosis of multiple sclerosis: a consensus state-
ment. Arch Neurol 2005; 62: 865–70.
Heussinger N, Kontopantelis E, Gburek-Augustat J, Jenke A,
Vollrath G, Korinthenberg R, et al. Oligoclonal bands predict mul-
tiple sclerosis in children with optic neuritis. Ann Neurol 2015; 77:
1076–82.
Huss AM, Halbgebauer S, Ockl P, Trebst C, Spreer A, Borisow N,
et al. Importance of cerebrospinal fluid analysis in the era of
McDonald 2010 criteria: a German-Austrian retrospective multicen-
ter study in patients with a clinically isolated syndrome. J Neurol
2016; 263: 2499–504.
McDonald WI, Compston A, Edan G, Goodkin D, Hartung HP,
Lublin FD, et al. Recommended diagnostic criteria for multiple scler-
osis: guidelines from the International Panel on the diagnosis of
multiple sclerosis. Ann Neurol 2001; 50: 121–7.
Montalban X, Tintore M, Swanton J, Barkhof F, Fazekas F, Filippi M,
et al. MRI criteria for MS in patients with clinically isolated syn-
dromes. Neurology 2010; 74: 427–34.
Polman CH, Reingold SC, Banwell B, Clanet M, Cohen JA, Filippi M,
et al. Diagnostic criteria for multiple sclerosis: 2010 revisions to the
McDonald criteria. Ann Neurol 2011; 69: 292–302.
Polman CH, Reingold SC, Edan G, Filippi M, Hartung HP, Kappos L,
et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the
‘McDonald Criteria’. Ann Neurol 2005; 58: 840–6.
Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers
GC, et al. New diagnostic criteria for multiple sclerosis: guidelines
for research protocols. Ann Neurol 1983; 13: 227–31.
Presslauer S, Milosavljevic D, Huebl W, Aboulenein-Djamshidian F,
Krugluger W, Deisenhammer F, et al. Validation of kappa free light
chains as a diagnostic biomarker in multiple sclerosis and clinically
isolated syndrome: a multicenter study. Mult Scler 2016; 22:
502–10.
Rovira A, Swanton J, Tintore M, Huerga E, Barkhof F, Filippi M,
et al. A single, early magnetic resonance imaging study in the diag-
nosis of multiple sclerosis. Arch Neurol 2009; 66: 587–92.
Ruet A, Arrambide G, Brochet B, Auger C, Simon E, Rovira A, et al.
Early predictors of multiple sclerosis after a typical clinically isolated
syndrome. Mult Scler 2014; 20: 1721–26.
Senel M, Tumani H, Lauda F, Presslauer S, Mojib-Yezdani R, Otto M,
et al. Cerebrospinal fluid immunoglobulin kappa light chain in clin-
ically isolated syndrome and multiple sclerosis. PLoS One 2014; 9:
e88680.
Swanton JK, Rovira A, Tintore M, Altmann DR, Barkhof F, Filippi M,
et al. MRI criteria for multiple sclerosis in patients presenting with
clinically isolated syndromes: a multicentre retrospective study.
Lancet Neurol 2007; 6: 677–86.
1084 | BRAIN 2018: 141; 1075–1084
Tintore M, Otero-Romero S, Rio J, Arrambide G, Pujal B, Tur C,
et al. Contribution of the symptomatic lesion in establishing MS
diagnosis and prognosis. Neurology 2016; 87: 1368–74.
Tintore M, Rovira A, Rio J, Otero-Romero S, Arrambide G, Tur C,
et al. Defining high, medium and low impact prognostic factors for
developing multiple sclerosis. Brain 2015; 138: 1863–74.
Tintore M, Rovira A, Rio J, Tur C, Pelayo R, Nos C, et al. Do
oligoclonal bands add information to MRI in first attacks of mul-
tiple sclerosis? Neurology 2008; 70: 1079–83.
G. Arrambide et al.
Voortman MM, Stojakovic T, Pirpamer L, Jehna M, Langkammer C,
Scharnagl H, et al. Prognostic value of free light chains lambda
and kappa in early multiple sclerosis. Mult Scler 2017; 23:
1496–505.
Zipoli V, Hakiki B, Portaccio E, Lolli F, Siracusa G, Giannini M,
et al. The contribution of cerebrospinal fluid oligoclonal bands
to the early diagnosis of multiple sclerosis. Mult Scler 2009; 15:
472–78.
Downloaded from https://academic.oup.com/brain/article/141/4/1075/4866204 by guest on 16 March 2024