The influence of anti-PRL autoantibodies on PRL measurements determined by immunoassays was investigated in 10 patients with anti-PRL autoantibodies. Four different immunoassay systems (two double-antibody radioimmunoassays (RIAs), a single-antibody RIA and an immunoradiometric assay (IRMA)) were examined. Total and free PRL were extracted from sera by precipitating gamma-globulin with polyethylene glycol with and without acidification, respectively. PRL values determined by direct measurement were compared with total PRL values. The proportion of free to total PRL levels determined by each immunoassay in sera with anti-PRL autoantibodies was significantly lower than in control sera. Values obtained by direct measurement of PRL (a routine assay procedure) in control sera were similar to total PRL values, whereas in sera with anti-PRL (a routine assay procedure) in control sera were similar to total PRL values, whereas in sera with anti-PRL autoantibodies the values varied from one immunoassay to the other. In sera with anti-PRL autoantibodies, double-antibody RIA 1, RIA 2 and single-antibody RIA 3 yielded values lower for PRL than for total PRL (52 +/- 15% in RIA 1, 40 +/- 8.8% in RIA 2, 40 +/- 14% in RIA 3), while PRL levels determined by IRMA were not significantly different (112 +/- 14%). Immunoglobulin G purified from serum with anti-PRL autoantibodies dose-dependently decreased the recovery of PRL assayed by the double-antibody technique, while it did not affect that by IRMA. These data suggest that the presence of anti-PRL autoantibodies gives variable results depending on the immunoassays used.(ABSTRACT TRUNCATED AT 250 WORDS)