Discordant results from "real-world" patient samples assayed for digoxin.

Background: Digoxin has a narrow therapeutic index and is primarily renally eliminated. To optimize dosing of digoxin, therapeutic drug monitoring has been important since assays became available in the 1970s. Immunoassays are not specific, and cross-reactivity with endogenous and exogenous compounds has been reported for more than 20 years. Interassay concordance has not been investigated in recent years in "real-world" patient samples.

Objective: To identify whether different digoxin immunoassays produce clinically different results in real-world situations, estimate the frequency of discordance, and determine whether an equation-based estimate compares well with digoxin immunoassays.

Methods: Plasma samples were sent to 2 accredited laboratories simultaneously and the digoxin results were compared. Results of immunoassays conducted using the Cedia DRI Digoxin Assay and the DGNA Digoxin Assay were compared with an equation-based estimate of plasma digoxin concentration.

Results: Thirty-six digoxin samples were assayed; in 39% of these, digoxin concentrations were discordant and different dosage adjustments would have followed. The presence of digoxin-like immunoreactive substances may explain some of this discordance. The mean of the equation-based result was similar to the immunoassay results, but marked variability was evident. The DGNA assay produced higher results on 24 samples; 9 higher values occurred with the DRI method.

Conclusions: Commercial digoxin immunoassays frequently produce clinically significant discordant results. The equation-based estimate does not appear to be an acceptable alternative to therapeutic drug monitoring. Immunoassay manufacturers should be required to improve assay performance by including real-world blood samples in development and clinicians should consider digoxin assay results warily.