Serum digoxin in the presence of Digibind: determination of digoxin by the Abbott AxSYM and Baxter Stratus II immunoassays by direct analysis without pretreatment of serum samples

^a Author for correspondence. Fax 503-494-8148; e-mail greent{at}ohsu.edu.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
We have reevaluated the feasibility of using direct immunochemical methods to track free digoxin in patients receiving Digibind^®. We report here that results obtained by the Stratus II and AxSYM immunoassays on patients receiving digoxin (without Digibind), digoxin-fortified serum samples supplemented with Digibind, and a digitoxic patient treated with Digibind, show no clinically significant biases. We conclude that useful free digoxin concentrations may be obtained for Digibind-treated patients using either the AxSYM or Stratus immunoassays without subjecting samples to ultrafiltration before analysis.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Digibind^® (Digoxin Immune FAB) given to patients who are digitoxic can cause substantial interferences in tracking therapeutic concentrations of digoxin depending on the type of assay methodology used. This is especially true in competitive immunoassays (1)(2)(3)(4) . One widely accepted method of circumventing this problem involves pretreatment of patient samples by ultrafiltration (5) . Ultrafiltration is, however, costly and labor-intensive, and it delays the turnaround of test results. It also yields lower free digoxin values relative to other well-established methods of tracking digoxin (1)(2) . Furthermore, there are problems in subjecting samples to ultrafiltration. Ultrafiltration does not, for example, account for loosely bound digoxin associated with albumin. This introduces an underestimate of the true concentration of available digoxin in patient sera. This problem is aggravated in that Digibind, although an avid binding agent of free and albumin-bound digoxin in serum samples, is rapidly cleared from patient sera. This introduces a "rebound" effect wherein the concentration of "free" digoxin rises sharply concomitant with clearance of Digibind (4)(5)(6) . Hence, it is important to gauge adequate treatment of digitoxic patients with Digibind to ensure that patients are not underdosed; at the same time, it is equally important to not overestimate the clearance of digoxin after Digibind therapy. Because digoxin bound to albumin comprises part of the available pool of digoxin in circulation, failure to account for this fraction of digoxin in serum samples may lead to underestimation of the true circulating pool of available digoxin. In addition, rigorous systematic studies of the adequacy of micropartition devices in separating free from bound digoxin, unlike studies with other lipophilic drugs, have not been done.

These observations, and difficulties in knowing in the laboratory which patients have received Digibind for treatment of digitoxicity, led us to reexamine the feasibility of monitoring free digoxin by direct immunoassay. We report here test results obtained by the Baxter Stratus II immunoassay with an alternative, albeit similar, immunochemical method released by Abbott Diagnostics on the AxSYM II analyzer. The Baxter Stratus II direct digoxin serum immunoassay correlates with values obtained by ultrafiltration (1)(2)(3)(7)(8) , albeit with as much as a 0.6 µg/L positive bias relative to results obtained by ultrafiltration (1) . We present evidence that clinically useful data can be obtained for Digibind-free and -fortified samples in both the Stratus II and AxSYM analyzer assays without the need for ultrafiltration.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Digibind was obtained from Burroughs Wellcome. Immunoassays for digoxin were obtained as reagent kits from Abbott and Baxter and used in accordance with the manufacturer's stated protocol on Abbott AxSYM and Baxter Stratus II analyzers. Matched serum specimens drawn from patients receiving digoxin, pooled serum to which a fixed concentration of digoxin and varying concentrations of Digibind had been added (see text), and serum from a digitoxic patient treated with Digibind were all analyzed on the Baxter Stratus II and Abbott AxSYM analyzers using 6-point calibrators obtained from each manufacturer. All specimens were analyzed as split samples on both instruments. Both instruments were also monitored daily with quality-control sera containing a high or low digoxin concentration to track and maintain stability of the assays throughout the course of this study.

The Stratus II immunoassay relies on the competition of alkaline phosphatase-conjugated digoxin with serum digoxin in binding to a solid phase immunomatrix filled with antibodies that can bind both types of antigens. Digoxin is quantitated through a dose–response curve constructed using a 6-point calibration. The Abbott AxSYM immunoassay utilizes micro enzyme particle immunoassay technology, which involves binding of digoxin to anti-digoxin-coated latex microparticles, entrapment of the digoxin-loaded latex particles within a matrix cartridge, passage of digoxin-labeled alkaline phosphatase through the cartridge to allow for enzyme binding to unfilled digoxin-binding sites, and subsequent washing and addition of substrate (methylumbelliferyl phosphate) to quantitate the digoxin initially presented to the latex particles. Digoxin concentrations are calculated from the fluorescent products generated as a result of substrate passage through the matrix cell and the dose–response curve constructed using a 6-point calibration of the system. Both methods utilize curve fitting algorithms to construct a dose–response curve, which is inversely proportionally to the concentration of digoxin present in test samples.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
In comparing test results obtained by direct immunochemical analysis on serum samples for free digoxin between the Stratus II and AxSYM analyzers, we first examined the analytical performance of the AxSYM because the performance of the Stratus has previously been evaluated extensively with regard to free digoxin analysis (1)(3)(7)(8)(9)(10) . Within-run CVs on the AxSYM for serum control digoxin samples were 7.7% (n = 10) at a mean value of 0.65 µg/L and 4.1% (n = 10) at a mean value of 2.96 µg/L, respectively. Day-to-day CVs measured over a 32-day interval averaged 6.1% at a mean value of 0.99 µg/L and 4.1% at a mean value of 2.71 µg/L.

To determine any interference attributable to digoxin-like immunoreactive factors, we assayed serum for digoxin on 20 digoxin-free infants (<6 months old), 5 renal transplant patients, and 5 liver transplant patients, both with no prior administration of digoxin. We obtained average digoxin concentrations of 0.22, 0.24, and 0.29 µg/L, respectively (range, 0–0.46 µg/L). Other serum samples (n = 10) of patients assayed at random who were neither neonates nor candidates for renal or liver transplants yielded digoxin values of 0 µg/L, suggesting that the AxSYM will yield occasional falsely increased test results when digoxin-like immunoreactive factors are present. The interference with digoxin-like immunoreactive factors appeared roughly comparable with test results reported by several other immunochemical methods of analysis, including the Stratus II (2)(9)(10)(11) . The linear range of the AxSYM immunoassay ranged from 0 to 8 µg/L.

A linear regression analysis correlating test results obtained on the AxSYM with those from the Stratus II immunoassay of patients (n = 55) not treated with Digibind or with Digibind present in assayed samples is shown in Fig. 1 . The calculated regression equation for samples lacking Digibind is y = 1.00x - 0.121 (Fig. 1 , r = 0.99; S_yx = 0.172; y = AxSYM; x = Stratus II). The difference in test results between the samples assayed is statistically insignificant (P 0.54). Additional assays were also made on serum samples to which a known concentration of digoxin was added (target value, 13 µg/L) that were subsequently titrated with increasing concentrations of Digibind to the point where no further free digoxin was detectable. These studies yielded a regression equation of y = 0.958x - 0.43; r = 0.997; S_yx = 0.333. As in the case of Digibind-free samples, we observed no significant statistical difference between the two immunoassay test results (P 0.89). Both the Digibind-free and Digibind-added serum samples yielded comparable, superimposable test results over the same test response range (Fig. 1 ). The concentration of Digibind required to completely neutralize stock serum samples made up to a target value of 13 µg/L digoxin corresponded to 15 µg/L digoxin-binding capacity (760 µg/L Digibind, based on the manufacturer's stated digoxin binding capacity).

View larger version (12K): [in this window] [in a new window]   Figure 1. Scattergram plot of free digoxin measured on the Abbott AxSYM vs Baxter Stratus II Analyzers in Digibind-free () and Digibind-added () serum samples. Arrow indicates free digoxin recovered on patient given Digibind 13 h before collection of sample. With the exception of this sample, and a 0 µg/L value obtained 2 h after Digibind treatment, all other Digibind-added samples were created by titrating stock 13 µg/L serum digoxin with progressively increasing concentrations of Digibind to the point where no free digoxin was detected (see Results). (——–), linear regression line of best fit of Digibind-free serum samples; (- - - -), linear regression line of best fit of Digibind-added serum samples.

In addition to digoxin specimens with added Digibind, three specimens were also obtained from a patient who presented digitoxic and was treated with Digibind in the cardiac care unit. This patient presented digitoxic with a serum value of 6 µg/L. Two hours after she was given Digibind, the patient's digoxin concentration was measured as 0 µg/L; 13 h after infusion it had risen to 2.7 µg/L (Fig. 1 , arrow). Linear regression analysis of the values obtained for the three specimens analyzed on this patient yielded a regression equation of y = 1.05x - 0.052 (r = 1.00) for test results of the AxSYM vs the Stratus II immunoassay, indicating a very close match to the dose–response curves previously seen with Digibind-free specimens as well as the regression equation for serum samples with added Digibind (Fig. 1 ).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
Many investigators have called attention to discrepancies in the estimation of digoxin caused by Digibind, depending on the methodology used (1)(2)(3)(4)(7)(8) . Rainey (4) examined the recovery of digoxin by different immunoassay techniques in survey specimens to which Digibind had been added to simulate the early and late phases of Digibind clearance. He reported a group mean of 1.2 ± 0.2 µg/L (n = 21) free digoxin for the Stratus, which should have had on a theoretical basis a free digoxin concentration of 0.8 µg/L (assuming that the average association constant between digoxin and Digibind is 10⁹). Whether this deviation between the calculated value and the value measured experimentally represented an overestimate is unclear. In a separate, independent, large-scale study by Hansell (12) , the Stratus immunoassay came within 0.1 µg/L of the theoretical target free digoxin concentration expected in Digibind-supplemented pooled sera samples sent to several laboratories for evaluation.

There is, nevertheless, evidence of systematic biases between test results obtained by ultrafiltration and the Stratus (1)(2)(3)(4) . Some investigators have suggested that biases between the Stratus and ultrafiltration technique might be attributed to overestimates by the Stratus in measuring protein-bound digoxin other than that sequestered to antigen-binding fragment (Fab) (2) , or to differences in the recovery of digoxin metabolites within the ultrafiltrate (1) . Banner et al. ((3)) proposed that ultrafiltration underestimates "available" free digoxin in that it fails to account for loosely bound digoxin associated with serum proteins retained with Fab within the retentate of ultrafiltrate-treated samples. On the other hand, Ujhelyi et al. (1) compared Stratus and Syva Emit immunoassay test results for free digoxin in digitoxic patients treated with Digibind to test results obtained by the Abbott Tdx fluorescence polarization immunoassay after ultrafiltration (FPIA-UF). They stated that the Stratus assay was a better predictor of free digoxin concentration than the Syva Emit assay but that both had considerable positive bias relative to FPIA-UF and that FPIA-UF "remains the assay of choice" in determining free digoxin during Digibind therapy.

The rationale for this conclusion is not clear. Ujhelyi et al. (1) claimed, for example, that their FPIA-UF assay constituted a reference method based on earlier studies by Hursting et al. (5) ; however, the method Ujhelyi et al. used to calibrate their instrument, using serum-based calibrators, differed substantially from the protocols used by Hursting et al. Hursting et al. encountered severe matrix problems when comparing test results by equilibrium dialysis and ultrafiltration (Figs. 2 and 3 in reference 5). They resorted to the use of correction factors in scaling their ultrafiltration test results to match those obtained by equilibrium dialysis, assuming that equilibrium dialysis more accurately tracked true free digoxin concentrations in the samples tested. Ujhelyi et al. made no comparable corrections in the workup of their serum samples. Indeed, they reported no matrix effects by FPIA-UF, an observation deviating sharply from that of Hursting et al. Furthermore, Ujhelyi et al. also processed their serum samples at centrifugation intervals exceeding two times that reported by Hursting et al., and at three times the centrifugal force. Whether these latter modifications biased the outcome of their test results with regard to the accuracy of their ultrafiltration assay is unknown. It is also of some note that the greatest biases between ultrafiltration and direct immunochemical methods reported by Ujhelyi et al. on patients given Digibind occurred roughly 24 h after its administration. This is consistent with reassociation of digoxin with serum proteins and a resulting underestimate of free digoxin in ultrafiltrates concomitant with retention of loosely bound serum digoxin within the retentate fraction of the micropartition device during the rebound phase after Digibind therapy.

Regarding systematic evaluation of ultrafiltrates for free drug recovery, Bowers et al. (13) have described criteria for systematically validating ultrafiltration free drug concentrations. These criteria have not been systematically applied toward evaluation of free digoxin concentrations by ultrafiltration, nor have the effects of centrifugation intervals, load volumes, and centrifugal forces with regard to the recovery of free digoxin been studied, as they have with other well-known lipophilic therapeutic drugs. In this regard, although many lipophilic drugs have been evaluated systematically for analysis by ultrafiltration (14)(15)(16)(17) , in some instances substantial changes have been noted in the recovery of free analyte, depending on the centrifugation interval and analyte concentration presented to the micropartition device (15) . Overall there are a number of reasons why further work needs to be done in defining standardized conditions for determination of free digoxin. Among the questions to be resolved regarding ultrafiltration is how to take into account variations in patient albumin concentrations secondary to renal failure or hepatic insufficiency which will directly affect recovery of digoxin in filtrate fractions.

Our data show that the AxSYM yields matching test results with the Stratus (Fig. 1 ) and, furthermore, that the dose–response curves for both methods in Digibind-free and Digibind-loaded serum samples overlap. Because of essentially identical correlation results in the presence or absence of Digibind, it is highly unlikely that either assay is interfered with by Digibind. The overlapping linear response curves thus not only affirm that Digibind does not directly interfere in the analytical performance of the Stratus immunoassay, as has previously been reported (2)(3)(4)(6)(7)(10) , they also indicate that the AxSYM immunoassay is also highly unlikely to experience substantial interference from Digibind. If Digibind interfered, it should have caused response curves concomitant with the presentation of Digibind to each immunoassay that deviated from those of the Digibind-free response curves, a fact not borne out experimentally.

The serial removal of digoxin with progressively increasing titers of Digibind (Fig. 1 , open triangles) and the overlap in dose–response curves seen in tracking digoxin in Digibind-free and Digibind-treated sera (Fig. 1 , filled vs open symbols) confirm that clinically useful data can be obtained when assaying serum samples containing Digibind by direct analysis using either the Stratus or AxSYM immunoassays. Data can be obtained without resort to ultrafiltration while reducing costs in expendables required in the preparation of ultrafiltrates, with reduced labor and with the elimination of protracted delays in the processing of digoxin samples. Although the Physician's Desk Reference (1998 edition) does not advise clinicians to monitor Digibind-treated patients for free digoxin (18) , this recommendation appears to be based on earlier findings showing serious limitations in immunoassays prone to interference from Digibind. Laboratories now have a choice among more advanced methods free of substantial interference in tracking digoxin concentrations. As noted by Hursting et al. (5) , rapid, simple, and more direct monitoring of Digibind-treated patients could provide "a timely appreciation" of the course of therapy in shortening the hospital stays of patients who have been treated for digitoxicity. On the basis of the data presented here, the Stratus II and AxSYM immunoassays appear suitable in this regard in monitoring the course of Digibind treatment.

	Footnotes

 
Division of Laboratory Medicine, L471, Department of Pathology, Oregon Health Sciences University, Portland, OR 97201.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Ujhelyi MR, Green PJ, Cummings DM, Robert S, Vlasses PH, Zarowitz BJ. Determination of free serum digoxin concentrations in digoxin toxic patients after administration of digoxin Fab antibodies. Ther Drug Monit 1992;14:147-154. [ISI][Medline] [Order article via Infotrieve]
2. Dodds HM, Norris RL, Johnson AG, Pond SM. Evaluation and comparison of the TDx II, Stratus, and OPUS digoxin assays. Ther Drug Monit 1995;17:68-74. [ISI][Medline] [Order article via Infotrieve]
3. Banner W, Jr, Bach P, Burk B, Freestone S, Gooch WM. Influence of assay methods on serum concentrations of digoxin during Fab fragment treatment. J Toxicol Clin Toxicol 1992;30:259-267. [ISI][Medline] [Order article via Infotrieve]
4. Rainey PM. Effects of digoxin immune Fab (ovine) on digoxin immunoassays. Am J Clin Pathol 1989;92:779-786. [ISI][Medline] [Order article via Infotrieve]
5. Hursting MJ, Raisys VA, Opheim KE, Bell JL, Trobaugh GB, Smith TW. Determination of free digoxin concentrations in serum monitoring Fab treatment of digoxin overdose. Clin Chem 1987;33:1652-1655. [Abstract/Free Full Text]
6. Berkovitch M, Manjapra RA, Gerace R, Verjee Z, McGuigan M, Whyte H, Koren G. Acute digoxin overdose in a newborn with renal failure: use of digoxin immune Fab and peritoneal dialysis. Ther Drug Monit 1994;16:531-533. [ISI][Medline] [Order article via Infotrieve]
7. Honda SA, Rios CN, Murakami L, Morita T, Scottolini AG, Bhagavan NV. Problems in determining levels of free digoxin in patients treated with digoxin immune Fab. J Clin Lab Anal 1995;9:407-412. [ISI][Medline] [Order article via Infotrieve]
8. Bizzaro N, Finco B, Milani L. Fab treatment in acute digitalis intoxication: reliability of serum digoxin determination with the Stratus system. Clin Chim Acta 1993;217:225-227. [ISI][Medline] [Order article via Infotrieve]
9. Ray JE, Crisan D, Howrie DL. Digoxin-like immunoreactivity in serum from neonates and infants reduced by centrifugal ultrafiltration and fluorescence polarization immunoassay. Clin Chem 1991;37:94-98. [Abstract/Free Full Text]
10. Jiang F, Wilhite TR, Smith CH, Landt M. A new digoxin immunoassay substantially free of interference by digoxin immunoreactive factor. Ther Drug Monit 1995;17:184-188. [ISI][Medline] [Order article via Infotrieve]
11. Crossey MJ, Dasgupta A. Effects of digoxinlike immunoreactive substances and digoxin FAB antibodies on the new digoxin microparticle enzyme immunoassay. Ther Drug Monit 1997;19:185-190. [ISI][Medline] [Order article via Infotrieve]
12. Hansell JR. Effect of therapeutic digoxin antibodies on digoxin assays. Arch Pathol Lab Med 1989;113:1259-1262. [ISI][Medline] [Order article via Infotrieve]
13. Bowers WF, Fulton S, Thompson J. Ultrafiltration vs equilibrium dialysis for determination of free fraction. Clin Pharmokinet 1984;(Suppl 1):49-60.
14. Barre J, Didey F, Delion F, Tillement J. Problems in therapeutic drug monitoring: free drug level monitoring. Ther Drug Monit 1988;10:133-143. [ISI][Medline] [Order article via Infotrieve]
15. Hua L, Montoya LL, Forman LJ, Eggers CM, Barham CF, Delgado M. Determination of free valproic acid: evaluation of the Centrifree system and comparison between high-performance liquid chromatography and enzyme immunoassay. Ther Drug Monit 1992;14:513-521. [ISI][Medline] [Order article via Infotrieve]
16. Bock JL, Ben-Ezra J. Rapid measurement of free anticonvulsant drugs by direct liquid chromatography of serum ultrafiltrates. Clin Chem 1985;31:1884-1887. [Abstract/Free Full Text]
17. Zysset T, Zeugin T. Comparison of the Amicon Centrifree micropartition system with the Sartorius SM 13249E Centrisart I device to determine protein free phenytoin concentrations. Ther Drug Monit 1986;8:346-351. [ISI][Medline] [Order article via Infotrieve]
18. . Digibind^®. Physicians' Desk Reference, 52nd ed 1998:1017-1018 Medical Economics Data Montvale, NJ. .