Use of insulin immunoassays in clinical studies involving rapid-acting insulin analogues: Bi-insulin IRMA preliminary assessment

Abstract

In clinical studies involving rapid-acting analogues (RAAs), insulin immunoreactivity is frequently measured, including endogenous, regular insulin (RI) and RAA immunoreactivities. Such a procedure implies equivalent cross-reactivities of all insulins present in serum. Commercially available human insulin immunoassays have been widely used, but their limitations (including hemolysis and anti-insulin antibodies) were not fully investigated. The aims of our study were to compare cross-reactivities of RI and RAAs in buffer and in serum and to investigate insulin immunoassay pitfalls. Cross-reactivities were assessed using Bi-insulin IRMA (Schering Cis-Bio International) in phosphate-buffered saline (PBS)-1% bovine serum albumin (BSA) and in pools of sera spiked with RI and RAAs (lispro and aspart). To investigate the influence of hemolysis, a pool of sera spiked with RAA was mixed with a concentrated hemolysate (final hemoglobin concentration 10 g/L) and incubated for 3 h at room temperature. To determine interference by anti-insulin antibodies, insulin was removed using charcoal from 18 sera with anti-insulin antibodies and from 17 sera without detectable anti-insulin antibodies. These insulin-free samples were then spiked with RI and RAAs and the immunoreactivity was determined. Compared with buffer, cross-reactivity in serum for RI, lispro and aspart was lower (35%, 29% and 26% lower, respectively). Hemolysis degraded almost all RI and RAAs contained in the serum (>or=95%). Anti-insulin antibody interference was significant for RI and RAAs (p<or=0.004) and correlated with anti-insulin antibody level in the serum (p<or=0.001). In serum, RI and RAA cross-reactivities are slightly lower than in buffer. For RAA assessment, hemolysed samples should be discarded and anti-insulin antibodies should be removed from samples before immunoreactivity measurements.