Problems encountered in the immunoassay of alpha-foetoprotein in patients with hepatoma.

Immunoradiometric assays suffer from the phenomenon of paradoxical increase in count rate with increasing dilution of highly concentrated samples--the "hook-effect". This is particularly important for alpha-foetoprotein where in hepatoma may be thousands of times the upper limit of the range of immunoradiometric assays. Using a commercial kit method we have encountered some degree of hook effect in many patients with hepatoma. Significant hooking occurred after 1,000 to 2,000 micrograms/1 but the extent of the effect was sample dependent. Of the samples tested so far, only one patient produced a hook which extended into the standard curve counts range. This resulted in a measured values of 20 micrograms/1 (value on dilution: 100,000 micrograms/1). We recommend that this method should be used routinely in conjunction with a screening method e.g., counterimmunoelectrophoresis and that quantitation of very high values should be by radial immunodiffusion or a method with similar sensitivity.