CSF Immunoglobulin G May Be Useful for Multiple Sclerosis Diagnosis

When the 2017 McDonald criteria were released, they highlighted the potential usefulness of cerebrospinal fluid (CSF) immunoglobulin G (IgG) in the process of diagnosing multiple sclerosis (MS). Now, recent findings have confirmed its practicality in the absence of a gold standard.¹

Ultimately, data a Bayesian latent class analysis of data from 673 patients found an estimated sensitivity of 0.93 (95% Credible Interval [CrI], 0.89 to 0.96) and a specificity of 0.81 (95% CrI, 0.77 to 0.85). The CSF IgG analysis’s positive likelihood ratio was 5—indicative of a moderate increase in the probability of developing MS—while the negative likelihood ratio was 0.09, signifying a tenfold decrease in the likelihood of developing the disease.

The study’s “latent class analysis indicates good diagnostic properties of CSF IgG analysis for MS. This test could thus be useful, especially for patients who tested negative for the 2005 and 2010 McDonald criteria,” Jonathan Epstein, MD, MSc, and coauthors wrote. They added that the findings deserve to be confirmed prospectively.

“A prospective study focusing on a cohort of CIS and using the same methodology is required to properly assess the use of CSF IgG analysis in comparison with McDonald criteria,” they noted.

The study included data from every patient who underwent CSF IgG analysis for events that were suggestive of MS at Nancy University Hospital from 2008 to 2011. In total, 67.3% of patients were women. Overall, 47% (n = 316) of patients had oligoclonal bands and 47.1% (n = 317) of patients had CSF IgG oligoclonal bands and/or an IgG index >.70.

Epstein and colleagues noted that the CSF IgG positivity was “mainly driven” by the incidence of oligoclonal bands, as only a single patient in the cohort had an elevated IgG index without OBs. This, they wrote, confirms a low, and possibly nonexistent, the role of the IgG index in the 2017 McDonald criteria that is harmonized with the use of oligoclonal bands but not the index itself.

“Moreover, 11.6% patients who did not meet the McDonald 2005 or 2010 criteria (i.e., suspected of having MS even if not diagnosed because of negative criteria), were CSF IgG positive underlining the interest of a spinal tap in this setting,” they wrote.

At the end of the mean follow-up period of 2.65 (±2.35) years, 36.4% of patients were diagnosed with MS, and 14.4% have had a single clinical event suggestive of MS without meeting diagnosis criteria. The true prevalence estimate was 36% (95% CrI, 0.89 to 0.96). As such, although the latent variable distribution (the patient’s true condition of having or not having MS) was recognized by the model, the authors noted that the similarity of the model’s estimation of prevalence and the clinician diagnosis-derived prevalence is a good indication of the model’s consistency with clinical observations.

Notably, the sensitivity and specificity estimates for patients with events that were suggestive of the onset of remitting MS were similar to those for the entire sample—0.92 (95% CrI, 0.85 to 0.96) and 0.80 (95% CrI, 0.76 to 0.84), respectively—but was higher for patients with signs of the onset of progressive MS, at 0.95 (95% CrI, 0.84 to 0.99) and 0.88 (95% CrI, 0.78 to 0.94), respectively.

After a clinical event suggestive of MS, patients not meeting McDonald 2005 and 2010 criteria on MRI but with oligoclonal bands on CSF analysis might be considered at high risk of developing MS. This, the authors concluded, “is in line with the 2017 revision of the McDonald criteria, where [dissemination in space] of inflammation on clinical or MRI data is sufficient to make the diagnosis if CSF analysis is positive for IgG synthesis, even if [dissemination in time] is missing.”

Epstein and colleagues did acknowledge some limitations of the study. The first being that patients suspected of having MS who did not undergo CSF analysis were excluded and the group admits not being aware of the number of patients who may fall into that designation—though they assume the number is low due to the “nearly systematic use of spinal tap in this setting.” The group used CSF IgG analysis as a single test as if no MRI had been done, therefore the analysis of CSF-only properties was not adjusted. However, MRIs were taken into account, as indicated by the McDonald criteria.

REFERENCE

1. Gamraoui S, Mathey G, Debourveire M, et al. High performance of cerebrospinal fluid immunoglobulin G analysis for diagnosis of multiple sclerosis. J Neurol. 2019;266(4):902-909.

doi

: 10.1007/s00415-019-09212-4.