Comparison of double-antibody radioimmunoassay with Farr-technique radioimmunoassay and double-antibody enzyme immunoassay for alpha-fetoprotein.

We describe double-antibody procedures for determining alpha-fetoprotein in human serum. An equilibrium procedure can be done in 24 h with a sensitivity of at least 4 mug/liter and coefficient of variation of 5.5%. There are no interferences from normal human sera or sera with certain commonly seen chemical abnormalities. We also describe and discuss sequential procedures that range in sensitivity from 250 ng to 1mug/liter and require 24-48h incubation. The precise (mid-range)portion of the dose/response curve for sequential procedures can be shifted to higher or lower values by an adjustment of the time of preliminary incubation of antibody with unlabeled antigen. With a 37 degrees C incubation, a sequential procedure can be completed in 7 h. Sensitivity is 1 mug/liter, and coefficient of variation 8.0%. The relative merits of the above assay procedures are discussed. The double-antibody redioimmunoassay is twice as sensitive as the Farr procedure [J. Infect. Dis. 103, 239 (1958)], and it is free of the large and variable nonspecific precipitation that accompanies the precipitation of bound antigen with sodium sulfate solution. Double-antibody radioimmunoassay is superior to enzyme immunoassay in both sensitivity and precision.