Real-World Performance of CSF Kappa Free Light Chains in the 2024 McDonald Criteria

Kappa free light chains (KFLCs) in the cerebrospinal fluid (CSF) have a similar performance to CSF-restricted oligoclonal bands (OCB) for multiple sclerosis (MS) diagnosis. To help with implementation, we set out to resolve several remaining uncertainties: (1) performance in a real-world cohort and the 2024 McDonald criteria; (2) equivalence to OCB in the specific clinical scenario when demonstration of intrathecal immunoglobulin synthesis is essential for MS diagnosis; (3) which KFLC metric has the best diagnostic performance.

A retrospective study of 740 cases was conducted, categorised into three groups: MS (2024 McDonald criteria), other neuroinflammatory disorders and non-inflammatory groups. CSF and serum KFLC and albumin were assayed with immunoturbidimetry. OCB status was assessed using isoelectric focusing. Eight KFLC metrics were tested: CSF KFLC, KFLC index, three population-based models of the upper limit for the CSF/serum kappa quotient corrected for CSF/serum albumin quotient, and their corresponding intrathecal fractions.

The KFLC index and the KFLC intrathecal fraction performed as well as OCB; no cases were missed when KFLC was mandatory to achieve a MS diagnosis. Intrathecal fraction computation improved the performance of the population-based models.

In the setting of the 2024 McDonald criteria, KFLC metrics correcting for the CSF/serum albumin quotient were equivalent to OCBs. The intrathecal fraction provided no advantage over the KFLC index, which is simpler to compute. Importantly, the KFLC index can replace OCB when CSF positivity is essential for diagnosis. We provide an explanation for KFLC's comparable diagnostic performance despite its inability to identify CSF-only clones.