Qualitative assessment of intrathecal IgG synthesis by isoelectric focusing and immunodetection: Interlaboratory reproducibility and interobserver agreement

Abstract

Detection of intrathecal IgG synthesis is important in patients with suspected multiple sclerosis (MS). The recommended method for the detection of intrathecal synthesis of IgG is isoelectric focusing and immunodetection of oligoclonal bands. Recently “The Committee for European Concerted Action for Multiple Sclerosis” has recommended that the results of isoelectric focusing for the detection of intrathecal synthesis of oligoclonal bands should not only be stated as positive or negative for intrathecal synthesis; instead, the laboratory should provide a detailed description of the IgG pattern in both cerebrospinal fluid (CSF) and serum together with a conclusion concerning the presence of intrathecal synthesis. We studied the interlaboratory reproducibility and interobserver agreement of isoelectric focusing, and the recommended classification system for the assessment of intrathecal IgG synthesis, in two separate patient groups employing k statistics. We found a high degree of interlaboratory reproducibility (133 patients; k= 0.95±0.05) and interobserver agreement (356 patients; k=0.97$0.04) when the presence or absence of intrathecal IgG synthesis was assessed. The agreement was less pronounced, although still fully satisfactory, when the results were classified according to the detailed system by two laboratories (133 patients; k = 0.86±0.08) and two observers (356 patients; k=0.88±0.08). Two specific problems in the interpretation of isoelectric focusing patterns were identified: one related to the discrimination of a pattern with several closely spaced bands which may represent a monoclonal protein; the other related to determining whether systemic band synthesis was present. We conclude that isoelectric focusing and immunodetection is a very reproducible technique for the detection of intrathecal IgG synthesis. Well defined criteria and extensive standardization may, however, be necessary when more elaborate classification systems are employed.