The fractional error as a means of assessing agreement between two methods measuring the same variable.

During the past 15 years, digoxin has been routinely measured in serum by various immunoassays. Unfortunately, all of the currently available immunoassays exhibit various cross reactivities with drugs, digoxin metabolites, and endogenous digoxin-like immunoreactive substances (1, 2). We recently encountered several patients for whom we were unable to obtain a reportable digoxin concentration with the TD Digoxin II Assay’ (Abbott Laboratories, North Chicago, IL 60064) because of high serum background intensity. These patients were found to be receiving, concomitantly with digoxin, Inocor’ (amrinone lactate injection; Winthrop Laboratories, Aurora, Canada), a new cardiac inotropic agent that is excreted largely unmetabolined in the urine. We assayed for digoxin serial dilutions of samples from a patient who was taking both Inocor and digoxin, and also assayed the pharmaceutical preparation of Inocor for injection use (1 mL containing 5 mg ofamrinone lactate and 0.25 mg of sodium metabisulfite in water), in both the TD, and the aca (E. I. Du Pont de Nemours, Wilmington, DE). The results indicated a positive interference of Inocor with the mx digoxmnassay but not with the aca assay. The Inocor preparation gave an instrument readout of 2.08 nmolJL in the Th. Further to investigate the nature of this interference, we analyzed with the TD serum samples from six patients receiving only digoxin, having divided the samples into paired aliquots and supplementing one of each pair with Inocor. By attributing a “digoxin” value of 2.08 nmolIL to Inocor, we calculated expected readout values for the dilutions, based on digoxin concentrations of the unsupplemented samples. The results, tabulated below, suggest a direct additive effect of Inocor on the TD,, digoxin assay. All observed and expected results were within 2 CV of the assay.