Comparison of the determination of insulin by a monoclonal antibody-based immunofluorometric assay and by radioimmunoassay.

This paper describes an immunofluorometric assay (IFMA) for insulin and compares it with the classical radioimmunoassay (RIA). Monoclonal antibodies against insulin were produced and used to develop the IFMA. One, immobilized on microtiter plates, was used for capture, the other, labelled with Europium, was used as tracer antibody. The IFMA presents sensitivity to an amount of insulin of 3 pmol/l and acceptable values for intra- and interassay error. The IFMA presented superimposable curves for human insulin, Arg65/Gly66-split proinsulin and des-Lys64,Arg65, and no cross-reactivity with human proinsulin, Arg32/Glu33-split and des-Arg31,Arg32. The RIA showed 100% cross-reactivity with human proinsulin, 90% with Arg32/Glu33-split, 193% with Arg65/Gly66-split, 340% with des-Arg31,Arg32 and 170% with des-Lys64,Arg65. The assays were used to measure insulin in 300 serum samples from 50 subjects submitted to an oral glucose tolerance test (OGTT). Twenty were normal, 10 had impaired glucose tolerance and 20 non-insulin-dependent diabetes mellitus. The mean value (+/- SEM) obtained by IFMA was 166.7 +/- 12.1 pmol/l and the mean value obtained by RIA was 339.6 +/- 18.6, with a correlation of r = 0.80 (P < 0.01). Comparison of basal insulin levels of the different groups of individuals using IFMA or RIA led to the same conclusions. The area under the curve showed statistically significant differences only for the comparison between normal lean subjects and individuals with impaired glucose tolerance, when measured by RIA. Our data stress the importance of methodology definition when comparing insulin results.(ABSTRACT TRUNCATED AT 250 WORDS)