We addressed the issue of cross-reactivity of several commonly used phosphodiesterase inhibitors with radioimmunoassays for cyclic AMP, after we had observed a considerably high cross-reactivity with a noncommercial antibody. Theophylline, pentoxifylline, penthydroxifylline (BL 194), albifylline (HWA 138), torbafylline (HWA 448), A 80 2715, isobutyl methylxanthine, and the nonmethylxanthines amrinone and rolipram were dissolved in supplemented and boiled cell culture medium (RPMI 1640). These samples were assayed for apparent cyclic AMP in two different, commercially available radioimmunoassay kits (based on polyclonal antibodies), applying the nonacetylated protocol. Cross-reactivity was dose-dependent and nonlinear. Samples containing theophylline and amrinone exhibited the strongest cross-reactivity in assay A (NEN/DuPont): 3.0 +/- 0.5(-nM) and 2.4 +/- 1.1 (-nM) apparent cyclic AMP +/- SD at 1-nM spike, respectively. With the more sensitive assay B (Amersham), higher concentrations of apparent cyclic AMP were detected: from 7.9 +/- 0.4 nM (for albifylline) to 3.5 +/- 0.1 nM (for rolipram). Values were calculated from standard curves set up in the respective assay buffer, where culture medium controls resulted in 1.8 +/- 0.3 nM and 3.1 +/- 0.1 nM for assay A and B, respectively. The culture medium interference increased with rising cyclic AMP concentrations. Although comparatively low, this degree of cross-reactivity is relevant for in vitro experiments. Phosphodiesterase inhibitors are commonly administered at millimolar concentrations, and resulting cyclic AMP levels are often in the nanomolar range. Neglecting these findings may lead to falsely high readouts of cyclic AMP concentrations.