CSF findings
The most striking finding was a lack of CSF-restricted OCB in the majority of cases: 65% (52/80) of all patients were negative for OCB (N=51) or only transiently positive (N=1) (Table
1). Overall, OCB were absent in 66% (56/85) of samples. These findings are in stark contrast to MS, in which CSF-restricted OCB are present in ≥ 95% of patients, are thought to remain stable over the entire course of disease and are considered a diagnostic mainstay (
p < 0.000001 when compared to a reference work on OCB in MS [
17]). Of note, OCB disappeared later in the disease course in 1/2 initially OCB-positive patients with available follow-up data. 3/3 initially OCB-negative patients with follow-up data remained negative for OCB upon retesting.
Table 1
Intrathecal synthesis, cellular immune response and total protein concentrations in BCS and MS
Number of patients | 132 | | n.a. |
Number of samples | 146 | | n.a. |
Sex ratio (male to female) | 1:1.54 | | n.s. |
Median age at first lumbar puncture (range) | 32 years (4–60) | | n.d. |
Relapsing disease course | 27/102 (26.5%) | | n.d. |
CSF-restricted OCB, positive, samples | 34.1% (29/85) | | < 0.000001 |
CSF-restricted OCB, positive, patients | 35% (28/80) | | < 0.000001 |
CSF-restricted OCB, transiently positive, patients | 50% (1/2) | n.d. | n.a. |
Link’s IgG index, elevated, samples | 22.6% (7/31) | | < 0.000001 |
Link’s IgG index, elevated, patients | 25% (7/28) | | < 0.000001 |
CSF WCC, elevated, samples | 28.1% (27/96) | | < 0.000001 |
CSF WCC, median and range, if increased, samples | 27/μl (6–371) | | n.a. |
CSF TP, elevated, samples | 40.8% (31/76) | n.d. | n.a. |
CSF TP, median and range, if increased, samples | 75 (46–280) | n.d. | n.a. |
ACD, samples | 23.6% (17/72) | 1.9% (1/54) | < 0.0005 |
Immunosuppressive treatment may theoretically affect OCB frequency. However, the difference between BCS patients and MS in terms of OCB positivity was still highly significant after exclusion of all patients treated with immunosuppressants (IS) at the time of LP (post hoc subgroup analysis; corrected
p < 0.00001). Moreover, so far, an effect of IS on OCB positivity in MS has been shown only for natalizumab [
18,
19], which was not used in any of the patients analysed in this study.
The difference was equally marked when only patients with both radiological and histopathological evidence for BCS and thus particularly high diagnostic certainty were considered (OCB negative in 73.7% [14/19]).
Due to the retrospective nature of the study, MRI quality varied considerably among cases. To exclude the possibility that the low rate of OCB was due to unintentional inclusion of patients with diagnoses other than BCS, we defined (in a fashion blinded to OCB results) a subgroup of cases with available high-quality MRI depicting textbook-like BCS lesions (N = 25); all of these lesions showed three or more layers of demyelination. In this subgroup of cases with particularly high radiological confidence, the lack of OCB was even more pronounced than in the total cohort (OCB negative in 81.8% [18/22]; OCB not tested in 3).
Finally, to rule out any chance that the low OCB rate found in our study partially reflects differences in sensitivity between current and ‘historic’ methods used to detect CSF-restricted OCB, we compared the frequency of OCB in reports published during the past 10 years with that in the older reports. However, no significant difference in OCB rates was observed between the ‘current’ and the ‘historic’ subgroup, with an even lower rate in the ‘current’ subgroup (31 vs. 37%).
We were then interested in whether OCB may define distinct subgroups of BCS patients. However, no significant differences were found between OCB-positive and OCB-negative patients with regard to sex (male to female ratio 1:1.8 vs. 1:1.4), median age at onset (32 vs. 30 years), median age at LP (32 vs. 32 years), disease course at last follow-up (relapsing in 36 vs. 34%), number of Baló-like lesions (more than one in 48.1 vs. 55.6%), location of Baló-like lesions (presence of infratentorial and/or spinal cord lesions 3.4 vs. 13.7%) and co-existing non-Baló-like lesion (present in 48.1 vs. 55.6%). While more patients in the OCB-negative group had an unfavourable outcome (fatal or no/almost no recovery in 23.3 vs. 9.5% among OCB-positive patients; N = 64), the differences did not reach statistical significance. Moreover, as a potentially important bias, most fatal cases were reported in Chinese, Philippine and Japanese patients in the 1980s and early 1990s, when OCB were not yet generally determined in those countries. However, if only cases published during the past 20 years (1998 to 2017) were considered, again, more OCB-negative than OCB-positive patients had an unfavourable outcome (fatal or no/almost no recovery in 20.5 vs. 7.1%; N = 53; p=n.s.).
Conversely, no statistically significant differences in OCB frequency were found between male and female patients (31.3 vs. 37.5%), between children/juveniles and adults (50 vs. 32.9%), between patients with a relapsing and patients with a non-relapsing disease course at last follow-up (36 vs 34%), between LP in patients with a single Baló-like lesion and in those with multiple Baló-like lesions (36.8 vs 32.4%) or between LP in patients with additional, non-Baló-like lesions and in those without such lesions (36.4 vs 27.6%). A non-significant trend towards a lower frequency of OCB in patients with no/almost no recovery or fatal outcome compared with patients who had partial to full recovery at last follow-up was noted (20 vs. 37.3%; p=n.s.).
In accordance with the widespread lack of IgG OCB, Link’s IgG index was positive only in 22.6% (7/31) of all samples tested for that parameter. This is in contrast to the reference study in MS, which reported an elevated IgG index in 86.4% (51/59;
p < 0.000001) [
20].
The CSF WCC was elevated in 27/96 (28.1%) samples. This compares to a CSF pleocytosis rate in MS of 58.4% reported in a reference work on CSF in MS (
N = 267;
p < 0.000001) [
17]. Among patients with pleocytosis and available data, the median WCC was 27/μl (quartile range 14.75–32.75). The frequency of OCB was higher in samples with pleocytosis than in samples without pleocytosis (58.8 vs. 20%); conversely, the frequency of CSF pleocytosis was lower in OCB-negative patients than in OCB-positive patients (55.6 vs. 17.9%;
p = 0.013).
TP was elevated in 40.8% (31/76; median 75 mg/dl; quartile range 54–98.8) of all CSF samples. Most patients with elevated TP and available data were OCB-negative (77.8% or 14/18). A so-called ACD, i.e. an increase in the CSF total protein concentration without an accompanying increase in the number of CSF white cells, was found in 17/72 (23.6%) samples. ACD is typically found in Guillain-Barre syndrome but not usually in MS (ACD in 1/53 [1.9%] LPs in 43 patients with MS according to McDonald et al., 2001 [45 × RRMS, 8 × SPMS]; unpublished data, S.J.; p < 0.0005 vs. BCS). The majority of patients with ACD and available data (90.9% or 10/11) were OCB negative. CSF MBP concentration was determined in seven samples and was elevated in three (42.9%).
Data on OCB pattern, IgG/A/M CSF and serum concentrations and/or QIgG, albumin CSF and serum concentrations, lactate CSF concentrations and CSF/serum antibody indices for measles, rubella or varicella zoster virus (MRZ reaction) were reported only in single patients or not at all, preventing systematic analysis of these CSF parameters.
43.2% (35/81) of all CSF examinations showed neither pleocytosis nor elevated TP, and 15.1% (11/73) showed neither CSF pleocytosis or TP elevation nor intrathecal IgG synthesis. Of note, 12 further LP were classified as ‘normal’ by the authors, but no more detailed information was provided. While these LP could not be included in the statistical analysis, they suggest that the low rate of WCC pleocytosis, OCB positivity and IgG index elevation observed in our study may even underestimate the real lack in CSF abnormalities in BCS and, in consequence, the real degree of difference between BCS and MS in terms of CSF pathology.