Importance of cerebrospinal fluid analysis in the era of McDonald 2010 criteria: a German–Austrian retrospective multicenter study in patients with a clinically isolated syndrome

Multiple sclerosis (MS) is a chronic inflammatory disease mainly characterized by demyelination and axonal loss [1]. A formal diagnosis of MS is based on clinical and radiological findings with an increasing role of MRI examinations as established in the 2010 revision of the so-called McDonald diagnostic criteria [2]. Here, cerebrospinal fluid (CSF) analysis, and in particular the detection of intrathecal IgG oligoclonal bands (OCBs), is a supportive criterion for a diagnosis of primary progressive multiple sclerosis (PPMS) but not in the more common relapsing–remitting form (RRMS) [2]. Nevertheless, the importance of OCBs, especially in the context of differential diagnosis and misdiagnosis in MS, is shown in several studies [3‐7]. Most often, the disease starts with a single clinical attack which is termed clinically isolated syndrome (CIS) when the MRI criteria of dissemination in space and time are not fulfilled [1, 8]. Several biomarkers allowing the prediction of conversion from CIS to clinically definite multiple sclerosis (CDMS) have been suggested [9‐13]: Besides cerebrospinal demyelinating MRI lesions, especially OCBs restricted to the CSF of CIS patients are associated with a higher risk for conversion to CDMS independent of the baseline MRI results [14, 15]. Additionally, positive OCB predicts CDMS in children with optic neuritis [16]. However, most of these CSF studies were conducted in the context of previous diagnostic criteria, i.e., McDonald criteria 2005.

In this study, we aimed to evaluate the prevalence and predictive value of OCBs in the context of the revised McDonald criteria 2010 [2] and therefore retrospectively analyzed 406 patients with a first presentation suggestive of MS from ten centers in Germany and Austria. We compared MRI and OCB findings as well as the disease course and conversion to MS according to McDonald 2010 criteria over a follow-up time of up to 12 years (median 32 months).

We collected data from 406 patients, 277 (68 %) of whom were female. The mean age at clinical onset was 37 years (SD ± 12). At disease onset, 44/406 (11 %) patients fulfilled the McDonald 2010 criteria for MS [2], 137/406 (34 %) the Swanton criteria (dissemination in space) [17], 87/406 (21 %) the Montalban (dissemination in time) criteria [18], and 44/406 (11 %) had no brain lesions. Intrathecal IgG OCBs were detected in 351/406 (86 %) and in 310/362 (86 %) CIS patients. 229/310 (74 %) converted to MS clinically or on MRI according to the McDonald 2010 criteria during the follow-up period of up to 154 months (median 32 months). All patient characteristics, as well as MRI and OCB findings, are summarized in Table 1 and Fig. 1.

While the conversion rate (by clinical or MRI signs) in CIS patients showing intrathecal OCBs (310/362) was 74 % (229/310), it was 44 % (23/52) in those CIS patients with negative OCBs. In patients without brain lesions at first attack and presence of intrathecal OCBs (30/44), conversion rate to MS was 60 % (18/30), whereas it was only 21 % (3/14) in those without OCBs (Fig. 1), revealing a positive predictive value of 79 % and a likelihood ratio for conversion of 3.4 in this subset of patients.

The median conversion time to definite MS for CIS patients with positive OCBs was 25 months (95 % CI 21–34) compared to 47 months (95 % CI 36–85) in those patients without OCBs (Fig. 2). CIS patients with intrathecal OCBs were twice as likely to convert to definite MS as OCB-negative individuals (hazard ratio = 2.1, p = 0.0014).

In Table 2, sensitivity, specificity, positive (PPV) and negative (NPV) predictive values of MRI and CSF parameters concerning conversion from CIS to definite MS are summarized. Whereas OCBs show the highest sensitivity of 91 % and an NPV of 39 %, Montalban criteria and OCB yield the highest specificity of 95 % (same result is achieved with Montalban criteria alone) and Montalban criteria reveal the best PPV of 92 %.

Among the 44 MS patients satisfying the McDonald 2010 criteria at disease onset, follow-up disease activity (by clinical or MRI signs) was observed in 36 out of those 41 who were OCB positive (88 %), whereas only one of the three patients without OCBs developed a second clinical attack detected by clinical or MRI signs (Fig. 1).

The last revision of the McDonald diagnostic criteria for MS, dating from 2010, does not include CSF criteria for a diagnosis of RRMS, while OCBs may support a diagnosis of PPMS [2]. Nevertheless, studies according to the previous diagnostic criteria showed that OCB positivity in CIS patients is a predictor for conversion to CDMS in adults [11, 15, 20] and children [16] independent of other factors [15, 21]. We now investigated the prevalence and predictive value of OCBs in the revised McDonald criteria 2010 era. Only 11 % of our patients with a first manifestation suggestive of MS met the revised McDonald criteria at disease onset. Thus, for the majority of CIS patients, further information allowing estimating the risk to develop definite MS would be of (high) value. We found OCB positivity in 86 % of 362 CIS patients at clinical onset. Those patients were approximately twice as likely to convert to definite MS and within a shorter period of time as OCB-negative CIS patients. This is in concordance with results in other cohorts, mostly referring to the revised McDonald criteria 2005 [11, 15, 21]. Whereas MRI criteria (Barkhof, Swanton, and Montalban) showed a higher specificity, OCBs performed best concerning sensitivity and NPV. Combining MRI criteria and OCB did not add to sensitivity or specificity for the prediction of conversion to definite MS compared to MRI criteria alone, despite that both MRI and the presence of OCBs in CSF have been repeatedly shown to be independent predictive factors [15, 21]. In another subset of patients, i.e., those who do not show any brain lesions (11 % of all patients), OCB-positive individuals are three times more likely to develop definite MS than OCB-negative patients; hence OCBs are the only predictor of conversion in this subset of CIS patients.

Our data further underline the utility and importance of CSF diagnostics, especially the detection of OCBs. We thus continue to recommend the inclusion of OCBs in the diagnostic workup of patients under the differential diagnosis of an MS [3]. CIS patients showing a positive OCB finding are at a higher risk of developing a definite MS; particularly in CIS patients not showing lesions, OCBs are of great interest. This might be helpful for the clinician to decide whether or not a disease-modifying immunotherapy should be started.