J Chromatogr 1986;383:187-93. 21. Plum J, Daidrup T. Detection of digoxin, digitoxin, their cardioactive metabolites and derivatives by high-performance li-.
CLIN. CHEM. 34/12, 2547-2551 (1988)
Improved LiquidChromatographic/Immunoassayof Digoxinin Serum Judith A..Stoes’
and Steven J. Soldln
This HPLC/immunoassay procedure measures digoxin in serum with no interference from digoxin metabolites or digoxin-like factors. We used solid-phase (C18 and Diol) extraction, a C18 column with a tetrahydrofuran/water mobile phase, and final quantification by fluorescence polarization immunoassay. Deslanoside and gitoxigenin were used as the internal standard and the retention-time marker, respectively. The average CV for 30O-L samples at digoxin concentrations between 0.9 and 3.9 nmol/L was 9.3%. Minimum column lifetime with daily use was three months. We also compared results, for 49 samples from patients taking digoxin, obtained with the Abbott “TDx FPIA digoxin I” and the present procedure. Discrepancies between the two methods were substantial for 20% of the samples.
measurement for serum samples are those involving highperformance hiquid-chromatographic (HPLC) separation of the various species, followed by either immunoassay to quantify column eluate fractions (8, 11, 18-21) or postHPLC-colunin derivatization (22,23). We have developed a modified HPLC/immunoassay to meet our requirements for pediatric pharmacology studies and to serve as a reference procedure in our laboratory. Similar published procedures did not fit our needs because of their requirements for large (1-mL) sample volumes, the short column lifetimes involved, or the co-elution of dihydrodigoxin (DHD) and digoxin. We also compared results for 49 samples from patients taking digoxin, as analyzed with the Abbott TDx fluorescence polarization immunoassay (FPuL) and with our method.
Additional Keyphrases: digoxin-likefactors lism pediatric chemistry FPIAcompared
Materials and Methods
‘
‘
digoxin metabo-
‘
The question of accuracy in digoxin immunoassays has come under intense investigation owing to interferences from digoxin metabolites and endogenous digoxin-like factors (DLFs) (1-6). DLFs have been documented in several patient cohorts and the response of different immunoassays to DLFs is varied (7). For purposes of therapeutic drug monitoring, interference in digoxin immunoassays by DLFs in neonates rather than DLFs in patients with renallhepatic failure or in third-trimester pregnancy are probably of most concern. In contrast, the issue of metabolism is of more interest for the population at large and may be of considerably greater significance. There is evidence for extensive digoxin biotransformation, but there are conflicting reports of the difference between digoxin concentrations as measured in serum by immunoassay or by methods with improved specificity (8-14). Overestimation of digoxin concentrations in serum because of cross-reactivity with metabolites having low cardioactivity may contribute to the poor correlation between measured digoxin concentrations in serum and efficacy or toxicity of the drug (15-17). However, the influence of metabolism on the accuracy of routine digoxin analysis is still incompletely defined. Several research methods are capable of measuring digoxin only, i.e., without interference from metabohites or DLFs (8, 11, 18-26). To date, the only assays allowing this type of ‘Department of Clinical Biochemistry, University of Toronto, Toronto,Ontario, M5S 1A1 Canada. ‘Children’s Hospital National Medical Center ResearchFoundation, Washington,DC 20010. ‘Departments of Child Health & Development and Pathology, GeorgeWashingtonUniversity, Washington,DC 20037. dress correspondenceto this author Department of Laboratory Medicine, Children’s Hospital National Medical Center, 111 Michigan Avenue N.W., Washington,DC 20010. 5Nonstandard abbreviations:DLF, digoxin-like factor(s); s’pi&, fluorescencepolarization immunoassay; DIII), dihydrodigoxin;MD, digoxigemnmono-digitoxoside;BD, digoxigenin bis-digitoxoside; Rt, retentiontime; and LC/iiPIA, high-performance liquid chromatography/fluorescence polarization immunoassay. ReceivedMarch 24, 1988; acceptedSeptember 15, 1988.
Apparatus: The chromatographic apparatus consisted of a Model M6000A solvent-delivery system with a U6K irjector (Waters Chromatography Division, Milford, MA). A Model LC-55 TJV-variable detector and a Model 023 recorder (both from Perkin-Elmer Corp., Norwalk, CT) were used to monitor and record absorbance. We used a 4.6 x 100 mm column, packed with 3-pm particles of ODS2 (Chromatography SciencesCo., Inc., Montreal, Quebec, Canada) and protected with a precolumn in-line filter (2-pm pore size). For FPIA of digoxin we used the Abbott TDx according to the manufacturer’s recommendations (#{176}TDx Digoxin F’ immunoassay with trichioroacetic acid as the precipitant). Solid-phase extraction was carried out under reduced pressure with Diol and C18 Sep-Pak cartridges and a Sep-Pak Cartridge Rack (all from Waters Chromatography Division). Reagents: Tetrahydrofliran and 2-propanol were purchased from Burdick and Jackson, Muskegon, MI; “HPLC”grade acetonitrile and methanol from Caledon Laboratories Ltd., Georgetown, Ontario, Canada; and “Analar”-grade ZnSO4’ 71120 from BDH Chemicals, Toronto, Ontario, Canada. “Surfasil” siliconizing reagent (primary ingredient, dichlorooctamethyltetrasiloxane) and “BCA Protein Assay” reagent (27) were from Pierce Chemical Co., Rockford, IL. Glass tubes (12 x 75 mm) used for extraction were siiconized with Surfasil, 100 mL per liter of acetone. Fresh acetomtrile wash solution, 20 mL per liter of water, was made daily. The ZnSO4 solution (100 g/L) was filtered and stored at 4#{176}C; under these conditions it was usable for up to 60 days. Mobile phases were prepared daily by mixing appropriate volumes of organic solvent and distilled, deionized water after filtration through a 0.45-pm (pore size) Durapore filter (Millipore Corp., Milford, MA), then degassing by sonication for 30 mm. Standards: Digoxin and gitoxigenin were from Sigma Chemical Co., St. Louis, MO; digoxin metabohites from Atomergics Chemetals, Farmingdale, NY; and [12a-3H]digoxin from New England Nuclear, Cambridge, MA. Deslanoside was a generous gift from Sandoz Pharmaceuticals, Dorval, Quebec, Canada. Digoxin and other cardiac glycoside stock standards were dissolved and diluted in ethanol; stored at -20 #{176}C, they were CLINICAL CHEMISTRY, Vol. 34, No. 12, 1988 2547
stable for up to one year. Solutions for preparation of serum standards or for extraction were then diluted in phosphatebuffered (pH 7.4) isotonic saline containing not less than 50 mL of ethanol per liter to prevent adsorption of digoxin to glass, and stored for no longer than one week at 4#{176}C. Aliquots of these serum standards were stored at -20 #{176}C for up to one year. We used fresh frozen plasma with digoxin values by FPIA of
