The kappa index, calculated by dividing the cerebrospinal (CSF)/serum kappa free light chain (KFLC) ratio by the CSF/serum albumin ratio, is gaining increasing interest as an indirect marker of intrathecal activation of the humoral immune response. The demonstration of intrathecal synthesis is of particular relevance in the diagnostic work-up of suspected Multiple Sclerosis. However, the lack of consistent data on inter-laboratory agreement in CSF and serum KFLC measurements is one of the factors that hamper the use of kappa index in routine practice.
Aim of this study was to assess agreement in CSF and serum KFLC measurements and kappa index values across different laboratories.
Fifteen paired CSF and serum samples were analyzed in all participating laboratories (nr=8). Four centers used Binding Site instruments and assays, 3 centers used Siemens instruments and assays, and one center used a Siemens instrument and a Binding Site assay.
Absolute individual agreement between laboratories was calculated using a two-way mixed effects intraclass correlation coefficient (ICC). Cohen's kappa coefficient was used to measure inter-laboratory agreement on positive (≥5.8) kappa index values.
Within Binding Site laboratories, ICC for KFLC measurements was 0.96 (95%CI: 0.9-0.98) for CSF, 0.93 (95%CI: 0.63-0.98) for serum and 0.97 (95%CI: 0.94-0.99) for kappa index values. Within Siemens laboratories, ICC for KFLC measurements was 0.99 (95%CI: 0.97-100) for CSF, 0.93 (95%CI: 0.48-0.98) for serum and 0.95 (95%CI: 0.89-0.98) for kappa index values. ICC calculated for all laboratories was 0.93 (95%CI: 0.87-0.97) for CSF KFLC, 0.81 (95%CI: 0.53-0.93) for serum KFLC and 0.65 (95%CI: 0.43-0.84) for kappa index. Cohen's kappa coefficient for a positive kappa index was 0.89 across Binding Site laboratories, 0.70 across Siemens laboratories, and 0.77 across all laboratories.
There was an excellent agreement in CSF KFLC measurements and in kappa index values within laboratories using the same instrument and assay (Binding Site or Siemens), while serum KFLC measurements were less concordant. Agreement across all laboratories was decreased when including the laboratory using a Siemens instrument coupled with a Binding Site assay in the analyses. Concordance for a positive kappa index was substantial across all laboratories and within Siemens laboratories, and very good within Binding Site laboratories.