The interference with three serum digoxin assay methods of endogenous digoxin-like substance (EDLS) in the serum of low-birth-weight (LBW) infants was assessed. The serum from 5-mL blood samples obtained from each of 19 LBW infants was divided into four 0.5-mL portions. Each portion was spiked with 10 microL of a distilled water-ethanol solution with or without digoxin to produce final digoxin concentrations of 0 (control), 0.49, 0.98, or 1.96 ng/mL. Each portion in each patient was then analyzed by radioimmunoassay (RIA), fluorescence polarization immunoassay (FPIA), and radial partition immunoassay (RPIA) using the control portions to measure EDLS. Serum digoxin concentrations measured by each assay method were calculated by subtracting the EDLS concentrations in the control portions from the measured digoxin concentrations in the spiked samples. The mean +/- S.D. concentrations of EDLS measured by RIA and FPIA were 0.26 +/- 0.13 ng/mL and 0.33 +/- 0.16 ng/mL, respectively. Of the 19 control samples assayed by RPIA, 18 had EDLS concentrations less than 0.1 ng/mL; one sample reflected an apparent concentration of 0.11 ng/mL. Mean recovered digoxin concentrations by RIA at each spiked digoxin concentration were significantly different from those obtained by FPIA and RPIA. A low but significant correlation was noted between EDLS concentrations in serum samples assayed by RIA and FPIA. The RPIA method appears to be preferred over the RIA and FPIA methods used in this study for serum digoxin analysis in LBW infants because of acceptable accuracy and minimal interference by EDLS.