In 1987 we organized an External Quality Assessment (EQA) scheme for cyclosporine (CsA) assay supported by the National Research Council (CNR). In 1991 the CNR EQA joined with the French CsA interlaboratory program organized by the Service de Radiopharmacie et Radioanalyse, Lyon. At the present time, more than 150 laboratories from seven European countries participate in this survey. Since the inception of the scheme, 119 control samples (from pooled blood of transplanted patients) have been prepared and mailed to the laboratories. During the whole EQA period, a trend toward the use of specific and nonradioactive techniques has been observed. It appears that, in the 1992 cycle, 83% of the collected results have been produced by methods (high-performance liquid chromatography [HPLC] or specific immunoassays) developed to assay the native molecule of CsA without its metabolites; the nonradioactive techniques (HPLC, fluorescence polarization immunoassay [TDx FPIA], enzyme-multiplied immunoassay technique [EMIT]) accounted for 64% of total results. The automated techniques showed, as expected, a good between-laboratory precision (specific TDx FPIA coefficients of variation, [CV] = 9.0%; nonspecific TDx FPIA CV = 11.3%; EMIT CV = 10.6%). The reproducibility of Cyclo-Trac radioimmunoassays (RIAs) yielded adequate results (nonspecific RIA CV = 10.0%; specific RIA CV = 11.1%), while the precision of HPLC was slightly worse (CV = 14.1%). The comparison of measurements obtained by different analytical techniques was performed by regression analysis of all results produced by the different immunoassays versus HPLC assumed as reference. These data demonstrate that the immunoassays, even if using monoclonal-specific antibodies, produce results significantly higher than those obtained by HPLC (bias of EMIT, about +10%; bias of Cyclo-Trac RIA, about +20%, bias of specific FPIA, about +25%). Nonspecific methods (polyclonal FPIA and RIA Cyclo-Trac monoclonal nonspecific) measure, as expected, CsA concentrations 3 to 4 times higher than those found by specific methods. In particular we found that monoclonal nonspecific RIA gives values about 35% higher than those produced by polyclonal FPIA, thus confirming that more CsA metabolites are measured by monoclonal nonspecific RIA.