Insulin determination by specific and unspecific immunoassays in patients with insulinoma evaluated by the arterial...

The aim of this study was to compare insulin concentrations measured by a traditional radioimmunoassay (RIA) and a more specific enzyme-linked immunosorbent assay (ELISA) in blood samples obtained during the arterial stimulation and venous sampling (ASVS) test in patients with insulinoma.

Prospective case series.

In 14 patients with an insulinoma undergoing an ASVS test, blood samples were obtained from the hepatic vein at baseline and 60 s after calcium injection into an artery supplying the tumour and a control artery (supplying pancreatic tissue without tumour). A selective arterial calcium stimulation was performed in five additional patients without evidence for an insulinoma. We measured insulin by a traditional RIA and a specific immunoassay.

In patients with insulinoma, insulin concentrations increased between 2.3- and 24.2-fold (median 8.2-fold) when measured by RIA and between 7.3- and 59.4-fold (median 16) when measured by ELISA following calcium injection into the artery supplying the tumour. Following calcium injection into the control artery, insulin concentrations were 0.6 to 1.3 times (median 1.0) the baseline values by RIA and 0.5 to 2.5 times (median 1.1) the baseline values by ELISA. In patients without insulinoma, insulin concentrations increased following calcium stimulation between 0.7- and 2.1-fold (median 1.3-fold) when measured by RIA and between 0.6- and 4.7-fold (median 1.3-fold) when measured by ELISA.

When insulin is measured by specific immunoassays, a higher cut-off (i.e. five- to sixfold increase) rather than the traditional criterion of a twofold increase should be used to localise an insulinoma during the ASVS test. The increase in insulin concentrations following calcium stimulation is significantly higher when insulin is measured by a specific assay compared with results obtained with traditional RIAs.