A MULTICENTER STUDY OF COMMERCIALLY AVAIL ABLE ELISA KITS FOR THE DETERMINATION OF HUMAN-IGG SUBCLASSES

We evaluated two commercially available sandwich type Elisa procedures for the measurement of IgG subclasses in human serum. Assay kits from The Binding Site and the Central Laboratory of the Netherlands Red Cr oss Blood Transfusion Service were rested in six laboratories. The per formance of spectrophotometers, pipettes and dilutors were assessed a t each center. Within-run precision was estimated according to the Val tec method (Societe Francaise de Biologie Clinique). The overall coeff icient of variation ranged from 4 to 50% depending on subclass and kit . We also evaluated the IgG(2) and IgG(4) specificity using four sera containing a monoclonal IgG(2) or IgG(4) (kappa or lambda type). Using total IgG and immunoelectrophoresis as a comparative technique, IgG(2 ) kappa and IgG(4) kappa were both underestimated, IgG(2) lambda was o verestimated while IgG(4) lambda compared favorably. Polyclonal IgG su bclasses were frequently overestimated in these sera suggesting cross- reactions with either monoclonal IgG or other polyclonal IgG. Antigen excess was investigated and not encountered with either kit. Our resul ts demonstrate that these procedures are insufficiently accurate or pr ecise for routine clinical use.