Measuring the unbound concentration fails to resolve analytic interferences in digoxin immunoassays

Therapeutic drug monitoring of digoxin is well established in the clinical management of cardiac patients treated with the drug. Recently, target concentrations have been revised in patients with congestive heart failure to 0.5 to 0.8 microg/L, challenging the sensitivity limits of most immunoassays. These widely used methods are often criticized, particularly on specificity grounds resulting from interference from exogenous and endogenous sources. One solution to remove higher molecular weight interference has been to ultrafilter plasma samples before assaying. The present study included 261 patient digoxin samples and compared two commercial ultrafiltration devices (Centrifree and Micrcon) that share the same separation membrane (YM-30). The results showed widely discordant apparent unbound digoxin concentrations in the ultrafiltrate from these devices with a Deming regression line of Centrifree = 1.31 x Micrcon + 0.042 (95% confidence interval for slope of 1.237 to 1.391) and apparent unbound fractions ranging from 15% to 610% of the unfiltered plasma digoxin concentrations, suggesting that ultrafiltration did not resolve such interference issues. The concept of measuring lower unbound digoxin concentrations as a result of lower therapeutic range for total (bound plus unbound) digoxin will also render most immunoassays insensitive and inappropriately calibrated. There is a strong imperative to review digoxin monitoring practices in the light of current clinical imperatives for both specificity and sensitivity reasons.