Sample preparation and radioimmunoassay for circulating free and antibody-bound insulin concentrations in insulin-treated diabetics: a re-evaluation of methods.

Methods for blood sample preparation and radioimmunoassay of free and antibody-bound serum or plasma insulin concentrations, using polyethylene glycol (PEG) solution to precipitate bound insulin, have been evaluated and compared. Method A was rapid mixing of whole blood with PEG, followed by separation of bound insulin by centrifugation into Ficoll; method B was rapid PEG addition to whole blood, followed by centrifugation alone; method C was conventional PEG addition to thawed plasma or serum, followed by centrifugation to remove bound insulin. Method A was found to have acceptable performance and reproducibility; serum free insulin levels were not significantly different from those measured by the simpler method B. In samples from insulin-dependent diabetics, serum free insulin was significantly higher and bound insulin significantly lower in method A compared to the conventional method C. However, in samples from non-diabetics there was no significant difference between methods A and C. Spurious results were obtained by addition of PEG and assay of plasma and serum samples after either freezing and thawing or incubation at 37 degrees C for 2h, compared to PEG addition and assay of fresh plasma or serum samples. We conclude that conventional, delayed addition of PEG to plasma or serum may result in redistribution of free and bound insulin and that values more representative of in vivo blood insulin levels are obtained by rapid PEG addition methods.