The Effect of Proinsulin on the Immunoassay of Insulin and its Possible Relation to States of Hyperinsulinemia ∗ †

The measurement of insulin or insulin-like moieties by a variety of bioassay and immunologic procedures has resulted in the designation of such nonspecific varieties as typical and atypical insulin (1), bound and unbound insulin (2), suppressible and nonsuppressible insulin (3). Insulin determined by the radioactive immunoassay and referred to as immunoreactive insulin, IRI, has been judged by most to represent an accurate appraisal of absolute insulin levels in biologic fluids (4); as such, its biologic activity, although inferred, has never been entirely certain. In the immunoassay method, the immunospecificity of antigen to specific antibody is of necessity assumed. Accordingly, if the reaction between insulin-I¹³¹, insulin antiserum, and endogenous insulin is specific, then serial dilution of serum containing IRI will yield comparable assay measurements over the range of dilutions studied.

In this laboratory where a modification of the Yalow and Berson chromatographic method (4) or the dextran-coated charcoal technique (5) is used to measure radioimmunoassayable insulin, certain sera have consistently failed to show such agreement (6), and have resulted in progressively higher than predicted IRI values with serial dilution of the serum sample. In the main these sera have come from patients in whom the diagnosis of insulinoma was being entertained where unusually high levels were anticipated so that initial measurements of IRI were made at several dilutions. To date, sera from five patients with so-called reactive hypoglycemia and many samples from a patient with organic hyperinsulinism, an insulinoma, have demonstrated this phenomenon of greater than 100% increase in IRI over several-fold dilutions of serum, appropriate dilution eventually resulting in linear agreement (Fig. 1). Furthermore, when present, this phenomenon appears to be greater when absolute IRI values are higher (Fig. 2).