Kappa Free Light Chain Biomarkers Are Efficient for the Diagnosis of Multiple Sclerosis: A Large Multicenter Cohort Study

Abstract

Background and objectives: Kappa free light chains (KFLC) seem to efficiently diagnose MS. However, extensive cohort studies are lacking to establish consensus cut-offs, notably to rule out non-MS autoimmune CNS disorders. Our objectives were to (1) determine diagnostic performances of CSF KFLC, KFLC index, and KFLC intrathecal fraction (IF) threshold values that allow us to separate MS from different CNS disorder control populations and compare them with oligoclonal bands' (OCB) performances and (2) to identify independent factors associated with KFLC quantification in MS.

Methods: We conducted a retrospective multicenter study involving 13 French MS centers. Patients were included if they had a noninfectious and nontumoral CNS disorder, eligible data concerning CSF and serum KFLC, albumin, and OCB. Patients were classified into 4 groups according to their diagnosis: MS, clinically isolated syndrome (CIS), other inflammatory CNS disorders (OIND), and noninflammatory CNS disorder controls (NINDC).

Results: One thousand six hundred twenty-one patients were analyzed (675 MS, 90 CIS, 297 OIND, and 559 NINDC). KFLC index and KFLC IF had similar performances in diagnosing MS from nonselected controls and OIND (p = 0.123 and p = 0.991 for area under the curve [AUC] comparisons) and performed better than CSF KFLC (p < 0.001 for all AUC comparisons). A KFLC index of 8.92 best separated MS/CIS from the entire nonselected control population, with better performances than OCB (p < 0.001 for AUC comparison). A KFLC index of 11.56 best separated MS from OIND, with similar performances than OCB (p = 0.065). In the multivariate analysis model, female gender (p = 0.003), young age (p = 0.013), and evidence of disease activity (p < 0.001) were independent factors associated with high KFLC index values in patients with MS, whereas MS phenotype, immune-modifying treatment use at sampling, and the FLC analyzer type did not influence KFLC index.

Discussion: KFLC biomarkers are efficient tools to separate patients with MS from controls, even when compared with other patients with CNS autoimmune disorder. Given these results, we suggest using KFLC index or KFLC IF as a criterion to diagnose MS.

Classification of evidence: This study provides Class III evidence that KFLC index or IF can be used to differentiate patients with MS from nonselected controls and from patients with other autoimmune CNS disorders.

Figure 1. Comparison of the Different FLC Biomarkers Values Between Groups

CIS = clinically isolated syndrome; CSF KFLC = CSF kappa free light chains; CSF LFLC = CSF lambda free light chains; IF = intrathecal fraction; KFLC IF = kappa free light chains IF; KFLC index = kappa free light chains index; LFLC IF = lambda free light chains intrathecal fraction; LFLC index = lambda free light chains index; MS = multiple sclerosis; NINDC = noninflammatory neurologic disorder controls; ns = nonsignificant; OIND = other inflammatory neurologic disorder. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 2. FLC Biomarkers ROC Curve Analysis and Ability to Separate MS (±CIS) From Other Control Populations

Panel A shows diagnostic performances to separate MS and CIS from patients with NINDC and OIND (n = 1,621 for KFLC biomarkers and n = 811 for LFLC biomarkers). Panel B shows diagnostic performances to separate MS from patients with SC (n = 842 patients for KFLC biomarkers and n = 415 for LFLC biomarkers). Panel C shows diagnostic performances to separate MS from patients with NIND (n = 1,067 for KFLC biomarkers and n = 541 for LFLC biomarkers). Panel D shows diagnostic performances to separate MS from patients with OIND (n = 972 patients for KFLC biomarkers and n = 479 for LFLC biomarkers). CIS = clinically isolated syndrome; FLC = free light chain; IF = intrathecal fraction; KFLC = Kappa free light chain; LFLC = lambda free light chain; NINDS = noninflammatory CNS disorder control; OIND = other inflammatory CNS disorder.